Hearings of the
Committee on Rules
Subcommittee on Legislative & Budget Process
Thursday, May 2, 2002
The subcommittee met, pursuant to call, at 10:40 a.m. in Room H-313, The Capitol, Hon. Deborah Pryce [chairwoman of the subcommittee] presiding.
Present: Representatives Pryce, Hastings of Washington, Dreier, and Slaughter.
Ms. Pryce. Good morning, everyone. The subcommittee will come to order. Welcome to an original jurisdiction hearing of the Subcommittee on Legislative and Budget Process. This hearing will provide an opportunity to have a dialogue about the very important issue of how the Federal Government determines the various costs or revenues associated with proposed policy changes.
Today's session is part one of a two-part hearing, and we look forward to our second session, exactly one week from today, same time, same place.
While many may be familiar with the term "dynamic scoring," the aim of our hearing is to help clarify exactly what it is and what it is not and why we need to further examine the concept. To be certain, changing the way that revenue and expenditure estimates are currently done is no easy proposition, and we have a responsibility as policymakers to pursue the best possible estimating approach that we can find, which may mean including the use of more macroeconomic feedback effects.
Today we are delighted to hear from a panel of expert economists with impressive qualifications. They are David Malpass, who is Chief Global Economist for Bear Stearns, a leading securities and brokerage firm, and a former Staff Director for the Congress' Joint Economic Committee among other things; Stephen Entin, the President and Executive Director of the Institute for Research on the Economics of Taxation, who has also served as Deputy Assistant Secretary for Economic Policy at the Treasury Department under the Reagan administration; and last, but certainly not least, Peter Orszag, Senior Fellow at the Brookings Institution, who also served as Special Assistant for Economic Policy under President Clinton.
I believe that one of our witnesses, Mr. Malpass, is hung up on the tarmac somewhere. We will proceed without him, and hopefully he can join us at the end of the hearing. Thank you all for being here and for participating in this discussion.
With that, I turn to our ranking member, Ms. Slaughter, for her opening statement.
Ms. Slaughter. Thank you, Madam Chair. I understand that feeling of sitting on the tarmac.
I would very much like to believe that our nation can anticipate unlimited growth as far as the eye can see. I fervently wish that I could tell my constituents that we can fight an open-ended war, ensure the solvency of Medicare and Social Security, and provide unlimited tax cuts without jeopardizing the economic health of this Nation.
But I can't. The numbers don't add up. They have never added up. And I am continually surprised that in the current fiscal climate the committee would try to obscure the ever-encroaching reality that the economic policies of Congress will have real and negative consequences for millions of Americans.
I am still stunned that we have fallen so far so fast. In less than a year, a surplus of $5.6 trillion shrank by $4 trillion. This is the worst fiscal reversal in American history. If you set aside Social Security, the budget projections shifted from a surplus of $3.1 trillion to a deficit of more than $700 billion.
And now comes word that Federal tax revenues are coming in significantly lower than expected, despite assurances of an improving economy. According to the Washington Post, this fact is leading budget experts to nearly double the projected budget deficit for the current fiscal year. Individual tax receipts are running $40 billion below projections. And when all the receipts are collected, the projected deficit could be $30 billion to $70 billion higher.
Moreover, next week we will also be considering the President's request for a $27 billion supplemental spending bill, further exacerbating an ever-ballooning deficit.
The estimates make it clear that it will be much more difficult for the Nation to return to budget surpluses in the next few years and will place new pressure on programs that generations of Americans have paid into, such as Social Security and Medicare.
It has also been suggested that current Federal budgeting methods may significantly misrepresent the government's fiscal position. If you adjust official CBO projections to separate retirement trust funds from the rest of the budget and provide a more realistic estimate, it leaves a much bleaker picture, a deficit of over $3 trillion over the next decade.
To counter this picture, many of the Members of the majority of this body are advocating that CBO change the manner in which they estimate the budgetary effects of policy changes. They argue that lower taxes will boost economic growth, thereby eliminating the revenue loss that otherwise would be expected. By using this reasoning, deficit projections would simply disappear along with the political headache that overspending will pose.
The body should know better. We have been down this road before. During the early eighties proponents of this trickle-down theory insisted that huge increases in defense spending would be coupled with extensive tax cuts to yield a booming economy. The results were utterly predictable and completely devastating for many regions of the country, including mine. Deficits exploded, interest rates rose and thousands lost their jobs.
It sounds somewhat familiar today. Many regions of the country such as, I would point out, my own of Rochester, New York, have never fully recovered. It took hard choices to put our fiscal house in order, choices which cost many of my colleagues their seats in Congress.
It is tempting to justify spending policies with rosy scenarios claiming wondrous and dynamic but unproven long-term effects, but it is wholly irresponsible. Economics is a highly imprecise science. Estimating the nationwide effects of tax policies is especially fraught with error, as evidenced by macroeconomists' sorry history of incorrect forecasts.
I would also remind my colleagues that the American people in poll after poll have remained remarkably sensible about the Nation's budget. They want an honest, fiscally responsible budget plan that balances America's priorities, from support for the military, to education, to a prescription drug benefit for our seniors. They want a fiscally responsible budget that would protect the economy by paying down the national debt, by strengthening Social Security and Medicare and investing in the future.
That said, and I appreciate it is harsh, but I do look forward to hearing and reading the testimony of today's witnesses. Thank you very much.
Ms. Pryce. Mr. Chairman?
Mr. Dreier. Thank you very much, Madam Chairman, Ms. Slaughter. I appreciate the fact that both of you have put together what I think is a very important hearing. I wasn't planning to make an opening statement, but some things have been said that have led me to feel that I might want to make a couple of comments.
For starters, I think it is very important that as we look at the budget process itself that we take into consideration behavioral patterns. To me it is absolutely ridiculous that as we have looked at the scoring process that we have had in the past, we don't take into account economic projections.
As I listened to the statement from Ms. Slaughter, I thought about the years I have served here in the Congress. I was elected the day Ronald Reagan was elected President in 1980, and we faced very, very serious economic devastation leading up to 1980. It is true that we saw an increase in the size of both the deficit and the national debt during the decade of the 1980s. The most important thing to note is that we saw a doubling of the flow of revenues to the Federal Treasury in the wake of implementation of the Economic Recovery Tax Act of 1981, ERTA. And yes, we were spending huge sums of money in an attempt to rebuild our Nation's defense capability in the wake of the terrible tragedy that happened in the desert, in the Middle East and other problems for our national security. The by-product of that, of course, was victory in the Cold War, which we saw at the end of the decade of the 1980s.
During that decade, we also saw a dramatic increase in the level of expenditures on subventions, the area of spending which frankly in many ways did not bring about the kind of improvement that was necessary.
Having said all that, we still saw a doubling of the flow of revenues to the Federal Treasury based on the fact that we implemented the Economic Recovery Tax Act of 1981. And so I think that as we look at challenges, we are clearly always going to see economic problems that exist in regions of this country. We are never going to have an absolutely perfect economy in every part of the Nation.
But I will say that as we look at the policies that we have tried to pursue to encourage economic growth, yes, we have seen a shortfall in the revenue flow based on the reports that have just come out, obviously because we have seen a slowdown in the economy. I don't know what assurances we had that we were going to be seeing economic growth. We have clearly mitigated the impact of the economic slowdown that was brought on in large part by September 11, we have mitigated the negative impact in large part with the implementation of last year's tax package. Most economists, Democrats and Republicans alike, have acknowledged that.
So I think that again as we look at this scoring process, it is very important for us to take into consideration behavioral patterns and economic growth, which is, I believe, the best way for us to get back onto that path of trying to responsibly reduce the size of the national debt and trying to reduce our deficit.
But I believe a very important part of that is to allow hard-working Americans to keep some of their own money so that they can invest those dollars and, through the benefit of that investment, see that increase in the flow of revenues so that we can meet our national defense responsibilities and the other many different challenges that we face.
I look forward to hearing from our witnesses, Madam Chairman. I thank you very much for calling this hearing.
Ms. Pryce. Thank you, Mr. Chairman. We will get right to that. If we can hear from our panel, such as it exists now.
Mr. Entin, would you like to begin? Thank you.
STATEMENT OF STEPHEN ENTIN, PRESIDENT, INSTITUTE FOR RESEARCH ON THE ECONOMICS OF TAXATION
Mr. Entin. Thank you, Madam Chairman. My name is Stephen Entin, President of the Institute for Research on the Economics of Taxation. I would like to thank the House Rules Subcommittee on Legislative and Budget Process for holding this hearing on Federal revenue and expenditure estimating.
Chairman Pryce has asked how well the economic forecasting and revenue estimating models used by the Treasury, the Joint Committee on Taxation and the Congressional Budget Office perform. I think your immediate concern is to examine the accuracy of the information these agencies provide to the Congress in its work on budget and tax issues.
As my testimony discusses, the current forecasting models generally ignore the consequences of changes in Federal tax and spending policies on the economy, which distorts the estimates of the effects of the policy changes on the Federal budget. But there is another problem. If you would adopt a more accurate dynamic method for predicting the budgetary effects of policy proposals by taking account of their effects on the economy, you would improve Federal budget forecasting, but you would do more than that. In the process, these new methods of forecasting would inform policymakers of the economic consequences of proposed policy changes. It would help you to choose more wisely among the various possible policy initiatives. The result could be a far better Federal tax system and a level and structure of Federal outlays more conducive to a strong economy.
Federal Reserve Chairman Alan Greenspan testified on the forecasting issue at a joint hearing of the House and Senate Budget Committees on January 10, 1995. He observed, "I don't think anybody who has looked at this system in considerable detail would say that the official scoring procedures which we are employing at this particular stage are realistic in terms of making judgments of choosing policy A versus policy B."
My former boss and founder of IRET, Norman Ture, testified at the same hearing and stated, "One thing can certainly be said about the existing system; it certainly produces wrong estimates in the case of all but a few tax changes."
Andrew Lyon, currently Deputy Assistant Secretary for Tax Analysis at the Treasury and formerly Associate Professor of Economics at the University of Maryland, said in 1995 at a National Tax Association conference, "It is hard to imagine that an economist with the ability to model the general equilibrium effects of a policy change would prefer the answers from a partial equilibrium model. If one's goal is unbiased estimation of the budget, it would seem to be important that dynamic effects be included with equal weight to those efficiency effects that are presently considered. When the macroeconomic effects are larger than those behavioral effects that are presently considered, it would seem prudent that these macroeconomic effects be considered first."
Congress should not be making policy decisions in the dark, nor should its policy options be artificially restrained or manipulated by budget trickery, even if the delusion is self-imposed to force itself to adopt supposedly responsible budget behavior. This is especially the case when the current forecasting techniques actually favor bad economic policies over good and result in irresponsible economic stewardship that hurts the Nation. Rather, Congress should have the best information that it can obtain to enable it to determine and adopt the policies that will best serve the country.
Although most private sector forecasting models recognize that changes in tax rates or other features of the tax code affect the aggregate economy in various ways, government revenue estimators at the Treasury and the Joint Committee on Taxation deliberately ignore tax-driven effects on the aggregate economy when attempting to measure the revenue consequences of proposed tax changes. Government revenue estimators sometimes claim that their models have dynamic elements because they sometimes allow for modest changes in which products are produced or where income is earned or how income is spent or what tax avoidance behavior might be practiced. But these models assume that tax changes don't affect total production, total employment, total earnings, saving and investment, economic growth or any of the other features of the aggregate economy. The government prepared estimation models are therefore essentially static.
When tax changes improve production incentives at the margin, output and income rise, which expands the tax base and yields positive revenue feedbacks. When tax changes worsen marginal production incentives, they yield negative revenue feedbacks. By ignoring these effects, the Treasury and the Joint Tax Committee deliver highly misleading revenue estimates and highly misleading advice.
Tax reforms with large positive work and saving incentives such as an across-the-board cut in tax rates, a lower capital gains tax rate, faster capital cost recovery allowances and elimination of the alternative minimum tax and estate tax would have much smaller revenue costs than government estimators claim. Spending increases that have no incentive effects, or tax cuts that have no incentive effects (there are types that don't), or transfer payments that discourage work and saving would decrease GDP, not add to it, and would cost more than initially forecasted by government revenue estimators.
It makes a difference how you try to model the dynamic effects. You can't just lump all tax cuts together or all spending changes together. You have to look at them and determine what they do as you put them into your model.
Some private models of the economy use a neoclassical framework in which tax and spending changes are assumed to affect behavior by changing the price signals in the economy. The output of goods and services is determined by the quantities of labor and capital services that are offered and employed. Certain types of tax cuts lower the cost of hiring and increase the value of saving relative to consumption, or of investment relative to consumption and alter behavior. These types of tax cuts do have an aggregate economic effect.
People often lump tax changes that don't have these marginal effects together with those that do not have these marginal effects. This is typical in the older Keynesian approach. They assume that shifts the Tax Policy Act by changing disposable income of individuals and the after-tax cashflows of businesses in ways that are unrelated to additional work or additional investment. In other words, tax cuts in the Keynesian framework are assumed to work by giving people and businesses money to spend, and tax increases slow the economy by taking money away from them.
This ignores the government budget constraint. If you cut taxes and you don't cut spending, you have to borrow the money back unless the Federal Reserve steps in and buys the debt for you. That type of tax change has no initial income effect and should not be scored that way as improving the economy.
Other models paint a complex picture of the results of fiscal policy changes assuming various reactions by the Federal Reserve and based on various theories, some of them suspect, as to how shifts in monetary policy affect the credit markets, price expectations and so forth. These efforts result in scenarios that are so speculative that they do more harm than good. The Federal Reserve should not automatically be assumed to lean against every fiscal policy change and trying to thwart any effort by the Congress to increase real growth. I don't think Federal Reserve officials behave that way and, if they do, they certainly shouldn't.
The several approaches to modeling the economy can yield very different predictions about the consequences of a specific change in Federal taxes and spending. The Joint Committee on Taxation and the Treasury often point to these differences and uncertainties as an excuse for not trying to factor any macroeconomic responses into their revenue estimates, as if the staffs of the Committee and the Treasury cannot exercise any independent judgment as to the relative merits of the differing theories and methods.
I think they know better than that, and I think they can do a good job if they focus on the models that try to track behavior changes from cause to effect very realistically. I think they can test various models one against the other based on historical data. I think if they get busy, they can do a good job for us, and I hope they will do so.
I want to make it clear that you are really dealing with two issues. One is the initial baseline forecast of what things are going to look like under current law. As Ms. Slaughter pointed out, these things can change very radically as the economy shifts. You can get wildly different predictions of budget surpluses and deficits from one year to the next or over a 2 or 3-year period. This is a problem that all economists face in projecting the economy.
But there is another issue, and that is, what will a policy change do given whatever the economy is going to do? What is the delta? What is the shift? What is the change from what would otherwise have happened? There, I think, we can be more useful and, if you want, more accurate because if we say this particular tax change is likely to increase GDP by 1 percent after 5 years. We don't know what GDP will be in 5 years, but whatever it is going to be, it will be 1 percent higher under this plan than under the other plan, or 1 percent lower, depending on what type of policy change you enact. I think that is an easier thing to forecast, and I think it is very important for you to have that kind of information.
That is where dynamic scoring really comes into play, which is in the economic and budget changes that would result from policy changes, not predicting a specific point estimate of where revenues will be 5 or 10 years out.
The Joint Committee and the Treasury have not gotten very far in their work on this effort. I don't think the Treasury has more than barely begun. The Joint Tax Committee has been doing work on this issue for about 5 years. They had a symposium some 4 years ago. I wish they had gotten further along, and I hope somehow they can be prodded to move a little faster and to make sure that whatever they do is made public so that it can be peer reviewed and that it can be tested against historical events so we can determine whether they are on the right track.
Thank you very much.
[The statement of Mr. Entin follows:]
******** INSERT 1-1 ********
Ms. Pryce. Thank you, Mr. Entin. We, the committee, appreciate your attendance here today and your testimony, and we will get back with some questions after we hear from Mr. Orszag.
STATEMENT OF PETER ORSZAG, SENIOR FELLOW, BROOKINGS INSTITUTION
Mr. Orszag. First, thank you very much for inviting me. I do think that periodically revisiting the scoring rules and the budget process is a very important thing to do, so I commend you for holding these types of hearings. I think maybe it would be appropriate just to underscore, since Mr. Entin quoted from Alan Greenspan and I have in the file I brought with me the transcript from that hearing. Let me just continue on with what Mr. Greenspan said about dynamic scoring:
"Full dynamic estimates of individual budget initiatives should be our goal. Unfortunately, the analytical tools required to achieve it are deficient. In fact, the goal ultimately may be unreachable."
He then continued:
"We must avoid resting key legislative decisions on controversial estimates of revenues and outlays. Should financial markets lose confidence in the integrity of our budget scoring procedures, the rise in inflation premiums and interest rates could more than offset any statistical difference between so-called static and more dynamic scoring."
Let me try to explain why that is Mr. Greenspan's judgment and why it is also my judgment that while the goal is a worthy one, at this point it would be an error to incorporate macroeconomic effects in individual budget scores. And to illustrate why, let's consider a recent example.
Following the September 11 attacks, there were proposals for large increases in public capital expenditures on infrastructure. If we were in a world of dynamic scoring, the cost of those proposals could have been offset by both the short run and long run effects of the public capital expenditures. So in the short run, the increase in government expenditures would boost aggregate demand, that would raise economic output and raise revenue and also reduce government transfer payments like unemployment insurance and other things. The net cost would then be lower than a macroeconomically static score would suggest.
In the long run there are various academic papers suggesting very high returns to public capital expenditures, for example, by reducing congestion in the Nation's airports and in other ways. So in the long run to the extent that these capital expenditures boosted output, they would raise revenue and if you netted out the cost from that extra revenue, it would look like the net cost of the proposals were much lower.
I should emphasize this point because a lot of times when people talk about dynamic scoring, they are only talking about it with regard to tax changes. But the exact same logic applies to spending changes. It is very important that if you do move to dynamic scoring, that it be applied to both sides, especially given the increasing use of the Tax Code to accomplish things that traditionally had been on the expenditure side. If we dynamically scored one and not the other, you would just exacerbate that tendency and that would be a very unwise move.
The challenges in undertaking that scoring, in both the short run and the long run, are described in my written testimony but they involve things like having to pick a model, having to, in the short run, assume something about what the Federal Reserve does, picking a parameter from the vast literature on the returns to public capital, with rates of return ranging from zero to as high as 200 percent in the published literature, and choosing a value from that literature that would be problematic.
And fundamentally, as my written testimony suggests, any time a very small movement in a parameter value, and "small" being measured relative to the credible published literature, moves the net cost of a proposal substantially, there is a big danger that the credibility and integrity of the scoring agencies would be undermined, because it would be left to their discretion to choose a value with very little empirical backing for choosing this value or that value. I could pick 15 or 20 percent returns or 25 percent returns and be well within the range of published estimates. There is very little guidance for picking 15 versus 25. If the difference between those two makes a huge difference in the net cost of the proposal, that is when you can run into a lot of danger in terms of the integrity.
One might ask yourself if -- up to now there has been very little partisan bickering about the scoring numbers themselves. You don't hear complaints that there were huge biases in this score or that score. With the dynamic scoring approach, those sorts of partisan complaints are much more likely, and in my opinion, that will endanger the integrity of the scoring process.
That is not to say that outside of the scoring process macroeconomic analyses are inappropriate. In fact, they are entirely appropriate. On that note, I quote in my written testimony from the analysis that the Congressional Budget Office did of last year's tax cut legislation, EGTRA. And I quote:
"Whether the tax cut will raise or lower real gross domestic product in the long run is unknown. But any effect is likely to be less than half of a percentage point in 2011."
A couple of things are of note here. First, the existence of the analysis itself. I think it is great that CBO does this sort of thing. They should continue to do that sort of thing. But note that they didn't assign a single point estimate to the long-term effect. They said there is a lot of uncertainty, it might be plus .5, it might be minus .5.
In the scoring process you don't have that kind of luxury of highlighting uncertainties involved and providing a range. You have to pick a number, and any time you have to pick a single number from what could be quite a broad range, you are raising dangers to the integrity of the entire process.
It is also important to note again, just to underscore, the agencies do do the type of macroeconomic analysis outside of the scoring process that could inform legislative debates about the pros and cons of various different pieces of legislation.
That then takes me to, well, what about just incorporating it directly into the scoring process? If they do it outside of the scoring process, anyway, why not just bring it into the scoring process?
My written testimony talks about the differences between those two, in addition to the fact that you need a single point estimate in a score, and you don't outside of it. There are timing constraints and resource costs to worry about.
Furthermore, in my opinion, the current budget process, if anything, is biased towards fiscal profligacy. The reason is that the baseline makes it look like there is too much budgetary resources available for either spending increases or tax cuts because it does not incorporate any adjustment to the alternative minimum tax, because it assumes the official sunsets in the various tax provisions, that expiring tax provisions are not extended, and because it assumes a dramatic decline in discretionary spending as a percentage of GDP or on a real per capita basis. And so it makes it look like there is more money on the table than there actually is.
Furthermore, the individual scores are not assigned debt service cost. So when you get an individual score back like on EGTRA last year, $1.35 trillion, that doesn't include the extra debt servicing costs that arise because we are reducing taxes and then the public debt is higher and therefore debt service costs are higher. EGTRA raised debt service costs by almost $400 billion on top of that $1.35 trillion.
So there are at least two substantial biases towards fiscal profligacy already built into the scoring process, and it would be particularly inappropriate to include dynamic scores without fixing these other biases, because we would then just be tilting -- to the extent that the dynamic score reduces the cost of the proposals, we would just be tilting even further in that direction and away from an unbiased approach.
Finally, it is worth pointing out that the results might not be all that some proponents of dynamic scoring think that they would be. For example, as I noted, CBO thinks that the long-term effect of last year's tax legislation would be plus or minus 0.5 percent as a percentage of GDP. That is a pretty minor effect. My colleagues Bill Gale and Samara Potter have just completed a rather exhaustive paper on the same piece of legislation and concluded that if anything, the effects are likely to be slightly negative in the long term. That is because the positive effects from better incentives to work and save from reduced marginal tax rates are offset by the negative effects of lower national saving, which reduces investment and reduces future output.
The point being that it may well turn out to be that all of this fuss is in exchange for very little change from many types of dynamic scoring and there would certainly be higher priorities that I would put in terms of reforming the budget scoring process rather than this specific step.
I would just mention a rather straightforward change that one could make is simply to include the debt servicing cost associated with each individual proposal in the individual score itself. That would be uncontroversial and straightforward to do, unlike the macroeconomic scoring that is the subject of this hearing.
Thank you.
[The statement of Mr. Orszag follows:]
******** INSERT 1-2 ********
Ms. Pryce. Thank you.
We have been joined by Doc Hastings. Is there anything you would like to say at this time?
Mr. Hastings of Washington. No. I certainly apologize for coming in late. I will just observe.
Ms. Pryce. We appreciate your presence. Thank you very much.
Mr. Entin, since we have one witness less than we had anticipated, is there anything you would like to say in response to Mr. Orszag? Would you both like to do a little back-and-forth for a couple of minutes? Or not?
Mr. Orszag. Sure.
Ms. Pryce. I think we would all benefit.
Mr. Entin. I mentioned in my testimony that we do need to look at government spending as well as tax changes in doing dynamic scoring. I would like to add we should also look at regulations, transfer payments and subsidies and such things. They all have effects that need to be brought into this.
I can understand why Mr. Greenspan would be nervous about going forward with dynamic scoring. He said the techniques haven't been developed. But we really haven't tried to develop them. That is part of the answer to his complaint. And if we tried to develop them along the lines of the economic theories that were popular 30 to 50 years ago, I think he would be scared because such things would predict bad outcomes and the markets might get spooked.
One thing it might predict is that higher spending would somehow raise interest rates and the bond market would get upset. Higher spending would raise interest rates because of the fear of inflation. The fear of inflation would be because the Federal Reserve might buy the additional Federal debt, and then the money supply might get out of hand, and then the inflation rate would go up and the bond prices would fall.
But Mr. Greenspan is in charge of the money supply and if he doesn't buy the added Federal bonds, none of that happens. So he is in effect saying, I would have to mess up in order for this bad outcome to happen, but I'm not going to mention that right now. I'm just going to mention the bad outcome. I think that is a little peculiar, because if he is afraid of that he can stop it by simply behaving himself. That analysis strikes me as being one we shouldn't wave around as a bogeyman against this kind of policy change.
The notion that government spending on infrastructure is going to boost the economy falls afoul of what I have brought up in terms of the government budget constraint. Thirty to 50 years ago it was very popular to say that a dollar increase in government spending would go right into aggregate demand and boost output by that much, and then the people who got paid to build whatever it was would then turn around and spend something too, and the people they brought from would spend something too, and that this ripple effect from spending would be stronger than the multiplier from tax cuts.
When the monetarists pointed out that the Federal Government has to pay for government spending, and borrow back what it gave out, this initial effect of boosting spending doesn't happen. So this multiplied expansion doesn't happen. If in fact you have a government spending program and it takes real resources, it is taking manpower and materials away from some other private sector use because we don't have 25 percent unemployment as in the Great Depression. The private sector investment that does not get done because the steel, computer chips, programmers and welders were put instead onto some government project might have had a higher return than the government project. It is the difference in the rates of return you have to look at.
So government spending should not be modeled as automatically adding to GDP. You have to do a cost-benefit analysis on it and see if its rates of return are markedly higher than those in the private sector or markedly lower, and only the difference should be counted as being a net increase or decrease in the economy.
The notion that national saving is affected by deficits is an old, old one and it is really time to put it to rest. When the government raises taxes to pay down the debt, people have less money to save. Generally speaking, when you raise taxes you do so in a way that discourages private saving and investment. So private sector saving tends to dip as government sector saving tends to rise. They largely cancel each other.
There is no major effect on interest rates due to the flow of funds or the net Federal borrowing requirement because when taxes go up, the first thing that the private sector cuts is its saving, not its spending. So national saving and interest rates do not change much, and that is not how the changes in policy should be analyzed.
There have been major studies over the years on the effect of deficits on interest rates. Some go back to the Civil War when we had deficits of record levels relative to the GDP, and commercial paper rates and railroad bond rates didn't budge. In World War II and World War I, it was the same thing. You do get an effect from tax changes on interest rates under some circumstances. Suppose, for example, Congress were to give 30 percent bonus depreciation, as was done in the recent stimulus bill, and if it were permanent rather than temporary so that people could count on getting it for the machines they put in place today and the replacements for those machines over time so that it is worth expanding the factory. Or consider some of the tax changes in the early nineteen-eighties that boosted the aftertax return to capital. In such cases, you would find, indeed, that physical investment became more profitable, rates of return went higher. Now, people can either buy physical assets directly or they can buy bonds to help somebody pay for a physical asset, or they can buy stock to help the company build a physical asset. Financial and real assets compete one with another, and if you raise the rate of return to capital in this society, real returns will go up both in the bonds and stocks and in the real machines. They are part and parcel of the same investment process.
Other than that (and, by the way, that is a good thing, because it means that you can attract money to fund the new investment), there is no independent credit market effect from the transitory deficit impact on the credit markets. That supposed credit market effect should not be the way these things are analyzed.
If you analyze things by looking at their impact on the incentives that govern real behavior, rather than these old ripple effects, and stick with that kind of a framework, most of these estimation problems simply disappear, and you narrow them down to a set of very tightly drawn behavioral responses that you can calculate without too much difficulty, as Mr. Lyon described in his testimony before the Budget Committee those years ago. Doing dynamic estimates that way, you have the same sort of tightly drawn rules that you have today, where we now assume that nothing happens. You would instead describe the buttons that could be pushed by the tax or spending bill, and note that some types of changes push the bottoms and some don't, and this is the set of rules we will apply to our revenue forecasts, and that we will stick to that, and we will treat all the Members' proposals evenly and fairly that way.
It depends on how you put the rule. The rule should be, figure out what really happens when you push these buttons, and stick to it, rather than assume that the buttons aren't connected to anything.
Ms. Pryce. Peter?
Mr. Orszag. Thank you. Where to start?
I guess let me sort of work in quasi-reverse order, first on the connection between fiscal policy and interest rates. Every major macroeconomic model, including the one used by the Federal Reserve, the one used by the Congressional Budget Office, the one used by the administration itself to project its baseline forecasts, the very models that Mr. Entin would like you to use to incorporate macroeconomic effects into tax legislation, show nontrivial changes in interest rates, long-term real interest rates, not short-term nominal interest rates, long-term real interest rates, from shifts in fiscal policy. Chairman Greenspan has testified repeatedly that he thinks there is a connection between interest rates and fiscal policy. There are economic theories under which there is not such a connection, but those theories are controversial and again not endorsed by the people who put together macroeconomic models and Chairman Greenspan.
In terms of whether there is a connection between national saving and fiscal policy, again there are controversial theories that suggest that there is no connection. Just to underscore what is involved in those theories, those theories assume that when the government borrows an extra dollar, everyone, including unborn people, look forward, know that they have to pay that dollar back in the future and therefore save more today. And if the unborn are not somehow doing that calculation, which of course would be quite a phenomenal feat, then their parents or grandparents take into account in deciding how much money to leave to their future descendants the fact that there will be higher debt.
This is, of course, a possibility, but it seems inconsistent with most human behavior. It is also inconsistent with the evidence. When we look at what happened during the 1980s and what happened during the 1990s, shifts in fiscal policy correspond on almost a one-to-one, not fully but on almost a one-to-one basis to shifts in national savings. That is also consistent with what again the administration and the Congressional Budget Office assume about what happens when you shift fiscal policy.
In terms of how to do analyses of public capital expenditures, I would agree, and my testimony underscores, that you need to compare the rates of return to public capital expenditure to the rates of return to private investment that it may displace, and that is the whole point. There is published literature from credible sources suggesting that public capital expenditures can have higher rates of return than private investment.
We can debate the credibility of those studies, but in a sense that is the whole point. If you are going to go down this road, there are lots of messy debates involving detailed econometrics that would be required to be adjudicated in some way with published studies suggesting 50, 60 and even one study suggesting 200 percent returns to public capital investment, which is much higher than the estimated rate of return to private investment.
I am not even getting into whether you would start to include the macroeconomic benefits from Head Start, from education, from health care. All of these things have -- to the extent that the advocates of dynamically scoring tax changes can point to some evidence suggesting that there are macroeconomic effects, the advocates of these other types of spending programs can point to equally credible evidence suggesting that they have large macroeconomic effects. I think that Chairman Greenspan's hesitation on this dimension along with the hesitation of the Congressional Budget Office reflects the fact that they don't want to be put in the middle of that ongoing debate among academic economists. From my experience in government, that seems like that is a reasonable position for them to take.
Ms. Pryce. Thank you very much, gentlemen. Mr. Malpass isn't present yet and I think if he doesn't appear we will just put his statement in the record. We do have his statement.
If you will permit us a few questions. It sounds like we all agree that tax reform, whatever we do, would affect people's behavior and the aggregate economy as well. But we don't account for these changes under current estimating conventions. Having the necessary tools is a critical component of making the best policy decisions, and some argue we don't have those yet. But is it preferable, as some would suggest, that we be precisely wrong rather than imprecisely right? If we know that this is going to be incorrect, why do we continue to do it that way?
Mr. Orszag. I think it is a very good question. I gave that quotation in my written testimony. I think that there are other considerations in the scoring process. Consider, for example, the following. Let's say that there is one estimate that may be in terms of its average expected effect somewhat more precise than another estimate, but that there is no controversy over the second estimate and a huge amount of controversy over this one, where changing that assumed parameter in this one can move it dramatically. It is not at all clear to me that we should be going this way rather than that way.
That in a sense is the whole point here, that while everyone agrees that there may well be macroeconomic effects, they don't agree on what they are. The recent tax legislation just underscores that. Some people, Marty Feldstein, Kevin Hassett, suggest you could get very large effects from last year's tax legislation. The Congressional Budget Office, Bill Gale, the paper that I have written, suggests it is going to be very small and it is possible to be negative. That kind of debate is a very messy one for the Joint Tax Committee and the Congressional Budget Office to be adjudicating in a real-time basis where you have got thousands of proposals that you have to be scoring and in my opinion would not be wise. That is not all that they would have to adjudicate.
Again, I think if this is done, it would have to be done for a spending proposal also. They would have to wade into the debate over what the returns to Head Start are in the long run, what the returns to education are, what the returns to smaller class sizes are. There are all sorts of potentially important macroeconomic effects from these types of proposals that are just the subject of a lot of controversy.
Mr. Entin. In many instances the different schools of thought will all say, yes, the direction is positive. Some will say by a billion dollars, some by 5 billion dollars, some by 10. If you want to be conservative, take something at the lower end of the range, but if they are all telling you this is beneficial, don't pretend that it isn't.
The point that you raise about there still being differences of opinion is an important one. We need to urge the Joint Tax Committee and the Treasury to do the research in advance as to what the real effects of these various changes are. They have already done a lot of research into various sorts of nonaggregate, nonmacro effects.
In Mr. Lyon's testimony before the Budget Committee, he pointed out that the tax increases that created the 36 and the 39.6 percent bracket were assumed by the Joint Tax Committee and the Treasury to have certain effects, such as that people would buy fewer corporate bonds and more municipal bonds. That was quite an effort. There were estimates of a little bit of movement and a lot of movement, but they settled on something. There were estimates that people would switch from stocks that had high dividends to stocks that had capital gains that they could defer because they didn't want to pay the higher rates on the dividends. They had to do a lot of econometric work to find out how much of that might occur. There were revenue estimates all over the lot, but they picked one.
There were other things that they did, but they didn't take account of the effect of the higher tax rate on the aftertax salaries of the people earning money in those brackets, such as doctors who might decide to retire a year early or work a little less, treat fewer clients, and perform fewer operations. That is no harder to do than the analysis of the sorts of things they did try to do.
As Lyon pointed out, the effect on the labor supply of those high earners on the total economy dwarfed in terms of budget consequences the little bits of changes between the muni bond and the corporate bond. They put so much effort into the stuff that didn't matter much, and very little effort into sorting through the literature to find out what the effects would be on the labor supply. They did nothing on whether or not the tax increase would affect the rates of return on investment by small business owners, which was another area that probably is as important as the labor effects.
RPTS O'ROURKE
DCMN BURRELL
[11:25 a.m.].
Mr. Entin. Mr. Lyon pointed out that if the estimators put some energy into sorting through the literature and came up in advance with known elasticities of substitution of labor for leisure, they could then refer to these response parameters whenever you ask a question at 3:00 in the morning in the middle of a budget debate. They would have the work done, and for example, be able to say that there is a 3 percent response for every 10 percent change in the rate of return. So, what did this amendment do? Well, it gives a 10 percent change in the rate of return, so you would have a 3 percent response. Click, click, click on the laptop computer and here is your answer. You cannot turn to an academic at MIT, Harvard or the University of Virginia at 3:00 in the morning and say, “Would you get up out of bed, go down to your mainframe and run this model for us?” That is not what we are asking. We are asking to get busy and do the work and narrow it down to a few response coefficients.
There is a wide range of estimates in the literature on the elasticities of virtually any inputs, but some of it is sensitive to the time period.
There were credible studies done in 1930 and 1940 that no one pays any attention to today because the data wasn't good and computers didn't exist back then, and the studies were wrong. They were credible when they were published. They are not credible now.
More recent data might lead to a different conclusion about some things. I get upset when CBO or the Joint Tax Committee surveys the whole literature back through the 1930s. They say, “Well, we are trying to figure out who said this and four articles said that.” But the articles from 1930 were just way off the wall and are no longer respectable, yet they just sort of lump them into the analysis.
Let me give you a notion of the change in thinking on deficits and interest rates. In the undergraduate textbooks as late as the 1960s and early 1970s, people were saying, “If I am saving this much each year and the government is borrowing this percentage of it, and if government borrowing goes up, and a greater percent of the current saving is being absorbed, then interest rates must rise.” In the early 1950s, the United States was the dominant economy in the world, Europe was devastated and Japan was devastated. We were it. So you would look at what was going on in the United States in measuring the effect of government on that credit market. And you would look at the flow of saving and the demand for saving just within the United States.
Well, two things have happened since then. First of all, we are an open world economy now. Many of the models the Joint Tax Committee tested on its tax reform symposium 4 years ago assumed a closed economy with no international capital flows. Well, that is highly unrealistic. You have to open the economy to the world credit market, not just the domestic credit market.
Second, we now know that when you talk about interest rates we are not talking only about the flow of funds this year versus the demand for funds this year. We are talking about the whole stock of accumulated saving over the last century in terms of the outstanding debt in the market today. It is not the amount of water flowing into or out of a pond, the stream going in versus the stream going out. It is the whole pond you have to look at. Most studies that try to look directly at deficits versus interest rates are using macro models that were built 30 years ago, assuming the old way of thinking. Test the relationship now against real data, and you come up with the following kind of analysis.
Even a large change in the Federal budget deficit would be only a small percent of the outstanding debt in the world. There are $80 trillion to $100 trillion dollars of financial instruments in the world credit markets that will be outstanding over the next decade. Assume an additional $1 trillion in added Federal deficit over a decade. It would add only about 1 percent to the stock of $100 trillion dollars in world debt.
Modern studies suggest that adding a percent to world debt would raise global interest rates by about a percent of themselves. To adjust the price of these bonds to encourage people to absorb another trillion dollars, you must move the price about a percent. That means the government bond rate might go from 6 percent to 6.06 percent, six basis points. Several studies have suggested that if you are looking only at the U.S. economy, that effect might amount to 18 basis points. If you had a world global credit market, it might only be six basis points.
These effects are trivial. The only time the deficit could upset the credit markets to any significant degree is if the Federal Reserve reacted in some bizarre way by trying to inflate the debt away. Then inflation would scare people and the bond market would react. But the effect of the deficit on the stock of debt outstanding is trivial. You have to look at the stock adjustment, not the flow of adjustments that were built into the major economic models in the 1960s. We are learning over time and, as we learn, we can narrow down these consequences and give you better predictions.
Ms. Pryce. Peter, I sense you want to respond.
Mr. Orszag. That would be wonderful. Thank you very much. Just two comments because I do think that this illustrates the nub of the argument. First with regard to whether there are ongoing controversies over -- let us just take the interest rate. Don't trust me, trust the Reagan administration, trust the first Bush administration. In testimony that I gave to the Senate Budget Committee in January, I quote from the Council of Economic Advisers from both administrations, again the first Bush administration only being about a decade ago. So unless things move quite dramatically it would be surprising to see a substantial shift then to the fact that fiscal policy did have a significant effect on interest rates.
I also note that under the theory that has been put forward for why they don't have a significant effect on interest rates, you would expect a significant effect between fiscal policy and the exchange rate. The changes in fiscal policy show up somewhere. If they don't show up in the interest rate, they will show up in the exchange rate. Either way there is consequences to be had. And I would just note that again it is not just Chairman Greenspan and, much more less importantly, me saying that there are interest rate effects but the Reagan and first Bush administrations themselves.
I think the comments about the 1990s are also illuminating. Increases in marginal tax rates that occurred in 1990 and 1993 are still the subject of ongoing debate in terms of what their effects are, let alone trying to estimate what they were before. I just published -- co-edited a volume that MIT Press put out called American Economic Policy in the 1990s, in which there is extensive discussion of the macroeconomic effects from the 1990 and 1993 Budget Acts. I don't even think its clear that the implication here is negative. If anything, at least among the people that were participating in this volume, the effects were seen as being positive again because they boost the national saving. But it doesn't matter whether you think they are positive or negative.
The point is there are extremely credible people on both sides of that issue and choosing one versus the other is going to be extremely problematic. I would worry if every time the leadership of the Joint Tax Committee or Congressional Budget Office changed we had a dramatic swing in the weight that we were putting on different factors in terms of determining macroeconomic effects, and that is precisely what would happen at this point given the current state of macroeconomic knowledge, unfortunately.
Ms. Pryce. Louise, since I could ask questions all day, why don't we alternate?
Ms. Slaughter. I have to read his testimony. But I was elected in 1986. We had a deficit about $390 million and we didn't expect it in our lifetime to see it. And we did think that the Budget Act of 1993 had a major impact and, as I pointed out, several of my colleagues lost their seats over it. I really think and I don't mean -- I hope you will take this in the spirit in which it is given -- but I think the economists are gamblers at heart. I really never understood how we can call it a science at all because it is so imprecise.
And I would like to add to the record one of the paragraphs from your testimony, Mr. Orszag, which really struck me. It is the one about Alan Auerbach says, quote, No government revenue estimator in my acquaintance would consciously provide a biased estimate."
But in many instances the uncertainty is so great that one honestly could report a number either twice or half the size of the estimate actually reported. Facing the threat of job loss and public criticism by Members of Congress and editorial writers each time an unfavorable estimate is reported, do we really expect estimators to flip a coin when they are unsure which is more accurate?
It may be that even if we have some information about the macroeconomic effects of policy proposals, reported estimates will actually be poorer if we insist this information be incorporated without standard errors in estimates used for budget scoring. You believe that quite strongly.
Mr. Orszag. If I could just, Professor Auerbach had been the Deputy Chief of Staff of the Joint Tax Committee, so he is not just an academic and also not one who is particularly partisan one way or the other but has some real world experience in these matters.
Ms. Slaughter. One thing that I think all of us learn as Members of Congress is that the dance of legislation that we do here is intended to please the markets and we have not been able to do that. But given the fact that right now despite the tax cuts -- and everybody is saying we are in the midst of recovery, in quotes, because I look around my district with job losses and it comes at us on almost a daily basis and the markets are taking a hammering -- does that mean they have no confidence in us at all? What is the connection between the market today and what the economic projections are? Everybody is saying we are poised for growth.
Mr. Entin. The markets are waiting for better corporate earnings to materialize. They are not sure when they will start, and it does affect the valuation of shares because people do tend to discount very distant years, so they have a short time frame. Earnings do come, but if they come 2 years later instead of 2 years earlier, if affects prices.
I don't think we should overestimate what government does to the economy. There are many types of tax and spending changes you can put in place that would all be regarded as more or less doing the same thing, perhaps largely nothing, to the economy. I don't think the markets are looking at you and blaming you for everything.
Ms. Slaughter. We would be sorry to hear that because that is what we are taught when we get here.
Mr. Entin. When I did my analysis of the tax bill last year, I remember saying this is going to do a little, but not much. I would have been happier with a different type of tax change which I think for the same dollar amount would have done better -- would have been better at promoting investment growth and employment.
It really makes a difference how you view these things. If you are going to do your analysis on whether or not the tax change changes the amount of capital that can be profitable, then you have one set of predictions as to the effect of the tax bill. If you simply assume you insert a certain amount of money into the economy and it reverberates around, that is not going to tell you whether that is a good tax change or not.
The markets, insofar as they watch the macroeconomic models, may not be looking at what the tax bill actually does. And if they look at what models say, if the models are generally wrong, then the markets are misled. I don't think that is what is happening. I think the markets know that the models are wrong. I think they are looking at the actual tax bill, and they are looking at what is going on with oil prices and they are looking at the fact that the tech sector hasn't gotten rid of its overcapacity as fast as we had hoped and are making a realistic judgment that the earnings rebound may be 6 to 12 months off and not happening in the current quarter.
And that leads me to the last point. I don't think any of the major models, the really big ones that had to be developed over 30 years or more, have built into them any realistic description of how what you do here today in Congress is going to affect the economy 1 to 5 years from now. So I wouldn't base my policy decisions on them. I would urge the Joint Committee and Treasury to start looking at some more modern ideas and do the studies that would show them how pushing button A or B affects the economy, and get those estimates lined up so they can use them.
What Lyon described in terms of the 1993 effort showed that the federal revenue estimators had to do in-house calculations on what the reaction would be as between corporate and municipal bonds, as between dividends versus capital gains, whether you would open a subchapter S corporation or C corporation, after the tax rates changed. They didn't turn to an econometric model in Pennsylvania or Massachusetts and ask what would the outcome be. These models have only the broadest assumptions about government behavior, and they calculate some general economic consequences, but they don't do the specific analysis that the revenue estimators need.
The estimators are going to have to do it in-house, and they are going to have to know what the techniques should be, and they are going to have to know the economics of it. I don't think they have done much yet to go down that road and that is what I am urging that you urge them to do. They should not take something currently on the shelf because there isn't something currently on the shelf that I would want to use.
But there are things in the academic community or models in the academic community that have gone part way in defining which connections there are between policy and behavior and have narrowed it down a great deal from this swarm of literature from 1930 onward, and have thrown out some of the nonsense.
When you look at that symposium that the Joint Tax Committee did in 1997, they had a number of the models operating as if there were no changes in international capital. Every one of those projection results should be thrown out. Remember what happened in Asia? And when the Russian ruble collapsed? And now the rebound, when they fixed what they did wrong? Capital does flow across boundaries. You can narrow down the assumption sets in the models so that the nonsense is thrown out. You can urge the technical people to look at the things that really matter and report back. Let them share it with the academics and then test it against history, and then let other people see their equations and work with the same data, the same equations, and see what they can do with it. Then see if you can get something that works well and then use it.
Ms. Slaughter. Dr. Orszag.
Mr. Orszag. Sure. Perhaps it is just worth noting a significant difference between the types of estimates that the agencies currently undertake and the ones that would be required under a macroeconomic effect type of estimation. The Joint Tax and the Office of Tax Analysis within the Department of Treasury have access to actually better resources than outside academics in order to estimate the individual effects in terms of labor supply -- I am sorry, in terms of shifting corporate status, et cetera, because they have access to individual tax data, not just the public use files that are very lagged but individual tax data, and you can try to tease out of those data responses when different tax changes occur.
Even in that area there are ongoing debates, but that is a wholly different level of activity and resources that are required relative to try to model the whole macroeconomy. You can't just simply do this without a model, and I would just say that there have been substantial efforts that have tried to improve the macroeconomic model, especially within the Federal Reserve Board, where they have replaced their macro model. The new model includes forward looking expectations and all the fancy gizmos that one would want to incorporate based on advances in economic thinking since the 1950s and 1960s.
And yet again, as I emphasized, Chairman Greenspan himself doesn't think we are there. So I don't want the subcommittee to think that there haven't been significant numbers. In my opinion there have been. The fact that we are not there, well, maybe that is because we are not trying hard enough. Maybe it is because the effort is so difficult that we would have to expend an unbelievably large amount of resources in order to get to an acceptable level, and I would just question whether that would be the best place to put our resources given the constraints on what we can do.
Ms. Pryce. We have been joined by David Malpass, and we are far past our hour. And so, David, if you could share your conclusions with us. Thank you so much for the trouble you took to get here, we are so happy to see you. You may proceed.
STATEMENT OF DAVID MALPASS, CHIEF GLOBAL ECONOMIST, BEAR, STEARNS & CO., INC.
Mr. Malpass. Thank you very much, Madam Chairman. I am sorry to be late coming down. The weather here apparently was insurmountable in getting in and I apologize to the committee and my colleagues. I will just state my conclusion.
As I look at the budget process, it is done in a way that patently does not come up with the right economic assumptions. For example, we have a scoring system in which a 20 percent across the board tax increase, a stunning increase would not conclude that that would change the economy at all. It would assume that no change had been made. And if you were to propose a major tax simplification, the scoring process would also assume that that would have no effect on the economy. That can't be the right answer.
My conclusion, as stated in my written statement: at some point, Congress has to decide whether and how it thinks the tax system, tax rates and the size of government impact economic growth. This is clearly a political issue as well as an economic issue. Congress has to recognize that the current system always comes out with the scoring answers that are wrong and that bias the system toward higher tax rates, greater tax complexity, and bigger government. So there ought to be a process to try to move toward a more accurate system of scoring.
I think if we don't do that -- and I am skeptical that we will make that change -- if we don't do that, I think we will remain mired in our current Tax Code, which in my mind significantly slows the economy and leaves us less competitive relative to other countries in the world.
Thank you.
[The statement of Mr. Malpass follows:]
******** COMMITTEE INSERT ********
Ms. Pryce. Thank you very much. All of the gentlemen are wonderfully credentialed and have added a lot of insight into the work that this committee and the Budget Committee and others have been doing.
Ms. Slaughter. I am going to have to be excused, and I want to thank all three of you. I am taking this testimony with me and I would like to call you up.
Ms. Pryce. And because we have overextended ourselves here, we will allow any statements to be submitted for the record. Please forward them to the Rules Committee.
This hearing will be continued at another session next week, where we will have the Director of the Congressional Budget Office, Chairman of the President's Council of Economic Advisors, and several Members of Congress. So you are welcome to come back to hear their perspectives.
Once again, it is a very exciting subject - one to which I, for one, am fairly new. It is all very interesting to me, and I appreciate the time you have taken to enlighten us. I am sorry that our time is up, but thank you for your efforts.
[Whereupon, at 11:45 a.m., the subcommittee was adjourned.]
