[Senate Hearing 114-122]
[From the U.S. Government Publishing Office]
S. Hrg. 114-122
DYNAMIC SCORING: HOW WILL IT AFFECT FISCAL POLICYMAKING?
=======================================================================
HEARING
before the
JOINT ECONOMIC COMMITTEE
CONGRESS OF THE UNITED STATES
ONE HUNDRED FOURTEENTH CONGRESS
FIRST SESSION
__________
JULY 28, 2015
__________
Printed for the use of the Joint Economic Committee
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
U.S. GOVERNMENT PUBLISHING OFFICE
95-802 WASHINGTON : 2015
-----------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Publishing
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800;
DC area (202) 512-1800 Fax: (202) 512-2104 Mail: Stop IDCC,
Washington, DC 20402-0001
JOINT ECONOMIC COMMITTEE
[Created pursuant to Sec. 5(a) of Public Law 304, 79th Congress]
SENATE HOUSE OF REPRESENTATIVES
Daniel Coats, Indiana, Chairman Kevin Brady, Texas, Vice Chairman
Mike Lee, Utah Justin Amash, Michigan
Tom Cotton, Arkansas Erik Paulsen, Minnesota
Ben Sasse, Nebraska Richard L. Hanna, New York
Ted Cruz, Texas David Schweikert, Arizona
Bill Cassidy, M.D., Louisiana Glenn Grothman, Wisconsin
Amy Klobuchar, Minnesota Carolyn B. Maloney, New York,
Robert P. Casey, Jr., Pennsylvania Ranking
Martin Heinrich, New Mexico John Delaney, Maryland
Gary C. Peters, Michigan Alma S. Adams, Ph.D., North
Carolina
Donald S. Beyer, Jr., Virginia
Viraj M. Mirani, Executive Director
Harry Gural, Democratic Staff Director
C O N T E N T S
----------
Opening Statements of Members
Hon. Daniel Coats, Chairman, a U.S. Senator from Indiana......... 1
Hon. Carolyn B. Maloney, Ranking Member, a U.S. Representative
from New York.................................................. 10
Witnesses
Hon. Phil Gramm, Ph.D., Former Chairman of the U.S. Senate
Committee on Banking, Housing, and Urban Affairs and Senior
Advisor, U.S. Policy Metrics, Washington, DC................... 3
Dr. Kevin A. Hassett, Director of Economic Policy Studies,
American Enterprise Institute, Washington, DC.................. 5
Dr. John W. Diamond, Edward A. and Hermena Hancock Kelly Fellow
in Public Finance, Baker Institute of Public Policy, Rice
University, Houston, TX........................................ 7
Mr. John L. Buckley, Former Chief of Staff to the Joint Committee
on Taxation, Washington, DC.................................... 8
Submissions for the Record
Prepared statement of Hon. Daniel Coats.......................... 32
Prepared statement of Hon. Kevin Brady........................... 32
Prepared statement of Hon. Carolyn B. Maloney.................... 33
Chart titled ``Corporate Tax Revenue as a Share of GDP Near
Historical Lows''.......................................... 36
Chart titled ``Total Tax Revenue as a Share of GDP''......... 37
Prepared statement of Hon. Phil Gramm............................ 38
Prepared statement of Dr. Kevin A. Hassett....................... 40
Prepared statement of Dr. John W. Diamond........................ 49
Prepared statement of Mr. John L. Buckley........................ 60
Report titled ``Dynamic Scoring and Infrastructure Spending'' by
Douglas Holtz-Eakin and Michael Mandel submitted by Vice
Chairman Kevin Brady........................................... 69
Report titled ``House `Dynamic Scoring' Rule Likely Will Mean
More Tax Cuts--Not More Information'' by Chye-Ching Huang and
Paul N. Van de Water submitted by Representative Carolyn B.
Maloney........................................................ 87
Questions for the Record for Dr. Diamond and Mr. Buckley and
Responses submitted by Representative Maloney.................. 92
DYNAMIC SCORING: HOW WILL IT AFFECT FISCAL POLICYMAKING?
----------
TUESDAY, JULY 28, 2015
Congress of the United States,
Joint Economic Committee,
Washington, DC.
The Committee met, pursuant to call, at 2:02 p.m. in Room
216 of the Hart Senate Office Building, the Honorable Dan
Coats, Chairman, and Kevin Brady, Vice Chairman, presiding.
Representatives present: Brady, Paulsen, Hanna, Schweikert,
Grothman, Maloney, Delaney, and Beyer.
Senators present: Coats, Cruz, Cassidy, Klobuchar, and
Peters.
Staff present: Cary Elliott, Connie Foster, Harry Gural,
Colleen Healy, Jason Kanter, David Logan, Kristine Michalson,
Viraj Mirani, Thomas Nicholas, Aaron Smith, Sue Sweet, and
Phoebe Wong.
OPENING STATEMENT OF HON. DANIEL COATS, CHAIRMAN, A U.S.
SENATOR FROM INDIANA
Chairman Coats. The Committee will come to order. Members
will be joining us. We are just finishing up on our caucus
luncheons, and I am told the House has some votes but they will
be drifting in also.
We want to start. We have got a terrific panel in front of
us, and we want to welcome our witnesses, including my former
colleague and very good friend, Former Senator Phil Gramm, who
I am not used to seeing on the other side of this dias. But I
thank all of our witnesses for being here today to discuss the
concept of dynamic scoring, a topic that has been much debated
since the House passed a rule earlier this year requiring the
Congressional Budget Office and the Joint Committee on Taxation
to use dynamic scoring when evaluating, ``major legislation.''
The Joint Committee on Taxation and Congressional Budget
Office have long provided lawmakers with estimates of spending
and revenue changes that would occur should a bill become law.
For decades, however, these scores, as they are known, have
largely ignored the largest driver of surpluses and deficits:
economic growth.
That is because the current method of estimation, known as
``static scoring,'' does not reflect the reality that the
economy can grow or contract as a result of public policy. Most
notably, it does not account for the massive effects a policy
can have on labor supply or private investment, two of the
largest drivers of the U.S. economy.
Ignoring these effects leaves lawmakers unable to debate
legislation with all available information at their disposal.
While dynamic scoring has been debated for decades, it is
no longer as it has been previously described, ``voodoo
economics.'' In fact, advances in computer technology and
economics have finally brought us from the question of ``Can
dynamic scoring be done?'' to the answer of ``Yes, and here's
how.''
We have the rare opportunity today to hear from those who
have been in the trenches of this debate as lawmakers,
Congressional staffers, and academics.
I would like now--well, I was going to recognize Ranking
Member Maloney for her opening statement, but let me turn to
the brief introduction of our witnesses today.
We are really privileged to have people here who have long-
time experience, and we are really looking forward to hearing
what their thoughts are as we go forward with the enormous
impact for decisions lawmakers have to make if we get this
right.
Senator Gramm served 6 years in the U.S. House of
Representatives, and 18 years in the United States Senate. His
legislative record includes landmark bills like the Gramm-Latta
budget, which reduced federal spending, rebuilt national
defense, and mandated the Reagan tax cut. And, the Gramm-Rudman
Act which placed the first binding constraints on federal
spending.
As Chairman of the Banking Committee, Senator Gramm steered
legislation modernizing banking, insurance, and securities law
which had been languishing in Congress for 60 years.
Those are but a few of the many substantive issues and
reforms that Senator Gramm introduced and brilliantly managed
to complete in his 24 years of service in both the House of
Representatives and the U.S. Senate.
Dr. Kevin Hassett is the State Farm James Q. Wilson Chair
in American Politics and culture, and Director of Economic
Policy Studies at the American Enterprise Institute. Before
joining AEI, Dr. Hassett was a senior economist at the Board of
Governors of the Federal Reserve System, and an Associate
Professor of Economics and Finance at Columbia Business School
in New York.
Dr. John Diamond is the Edward A. and Hermena Hancock Kelly
Fellow in Public Finance--boy, this is a mouthful here--at the
Baker Institute, Adjunct Professor of Economics at Rice
University, and CEO of Tax Policy Advisors LLC. His current
research focuses on the economic effects of corporate tax
reform, the economic and distributional effects of fundamental
tax reform in individual portfolio allocation in the 2000s, and
various other tax policy issues. He is co-author of ``The
Fundamental Tax Reform: Issues, Choices, and Implications,'' a
former editor for the National Tax Journal, and has served as
staff on the Joint Committee on Taxation from 2000 to 2004. So
welcome back, and sitting--there you are--sitting at the table,
rather than back here.
And finally, John Buckley has advised senior members of
Congress on tax legislation, and written extensively on the
subject. His career as a Congressional staffer spanned over 35
years, most recently serving as Chief Democrat Tax Counsel for
the House Ways and Means Committee until his retirement in
2010. He also served as Chief of Staff of the Joint Committee
on Taxation; and before that, Assistant Legislative Counsel on
the House Legislative Counsel's office. Off the Hill he has
been an Adjunct Professor at Georgetown University Law Center
for the last several years.
You know, I might take a little liberty here, Dr. Cassidy,
if there are a few opening remarks you want to make, I am happy
to do that. Otherwise, we will turn to our witnesses.
Dr. Cassidy. I am ready for the witnesses.
Chairman Coats. All right. Senator Gramm, you are on.
[The prepared statement of Chairman Coats appears in the
Submissions for the Record on page 32.]
STATEMENT OF HON. PHIL GRAMM, Ph.D., FORMER CHAIRMAN OF THE
U.S. SENATE COMMITTEE ON BANKING, HOUSING, AND URBAN AFFAIRS;
SENIOR ADVISOR, U.S. POLICY METRICS, WASHINGTON, DC
Dr. Gramm. Well, Mr. Chairman, first I am very happy to be
here. I am especially honored to testify before your committee.
We served for many years in the Senate together, and I
appreciate you affording me this opportunity.
I also need to mention that the Vice Chairman of the
Committee, Kevin Brady, is an old friend of mine from Texas.
Let me say that when I was in Congress I spent a lot of
time working on issues related to the economy and the budget.
And there may be people who have looked at more budget numbers
than me, and I am sure there are people who have looked at them
with a larger knowledge base than me, but I paid very close
attention to budgets when I was a Member of the House and the
Senate.
The one thing that I discovered that is irrefutable is that
the general condition of the economy overwhelms everything
else. Over and over again if you study these budget numbers,
you've got to conclude that we take actions to raise taxes, to
cut taxes, to increase spending, to reduce spending and almost
every member argues for every program they are for that it is
going to be good for the economy. But the bottom line is, the
changes in the economy swamp even the largest legislative
change that we make.
So obviously one of the objectives that I have always felt
that we needed to have was to find a way to take into account
in some manner the impact of our proposed policy changes on the
economy because most of the things that we undertake we claim
that we are trying to benefit the economy, whether it is an
increase in expenditures on some government program, or whether
it is cutting taxes, the objective that is presented in the
debate is almost always: this is going to be good for the
American economy, good for working people. And yet we, for all
practical purposes, have had no consistent ability to make even
a broad estimate of what the impact was going to be.
It seems to me that there are three conditions that ought
to be met for using dynamic scoring. And let me make it clear,
I am not talking about dynamic scoring as a substitute for
static scoring.
I think we do an excellent job in static scoring. I think
the Congressional Budget Office and the Joint Tax Committee
have gotten better and better at it, but I see it as a
supplement to static scoring because only when we bring the two
together can we look at the cost and benefits of various
policies.
So what are the three conditions that I believe should
exist for you to use dynamic scoring?
First, I think you have got to have a clear and consistent
economic theory that the policy is going to have a substantial
economic impact.
Secondly, you have got to have some evidence that it is
going to have an impact within the period that you are writing
the budget for, which is generally 10 years or less.
And finally, you need a base of information that shows that
in the past similar policies have produced empirical evidence
that would substantiate the claim that you are trying to make
about dynamic scoring.
I think probably one thing we would all agree on here, no
matter what our view is, is that the burden of proof ought to
fall on people that are arguing that we should use dynamic
scoring. Let me talk very briefly about two cases.
I want to talk about the Republican and President Clinton's
bipartisan agreement to balance the budget, cut the capital
gains tax rate, and increase the family tax credit.
The argument here is that the evidence is overwhelming. You
had an estimate by the Congressional Budget Office for the five
years after the proposal went into effect. When the five years
had ended, we had actually seen GDP go up by $2.4 trillion
above the CBO projection. That would be $4.7 trillion today.
It made up $8,609 over the five-year period for every man,
woman, and child in America. This was a significant policy
change. And revenues rose by over a trillion dollars during
this period. And so I think the evidence is pretty strong that
any effort to control spending as a means to balance the
budget, especially if it entails long-term policy changes like
entitlement reform, should be scored dynamically, and based on
the evidence of the Clinton era the scoring should be
substantial.
Finally, I think there is strong evidence to substantiate
the claim that a revenue-neutral tax reform proposal, if it
lowers marginal rates and eliminates inefficiency in the system
by the elimination of deductions and subsidies, that there is
evidence that that has produced strong economic results.
Everybody forgets that by 1988, when the full rate
reductions of tax reform kicked in, we were deep into the
recovery. It was already one of the longest recoveries of the
post-war era, and a third longer than the average recovery had
been in the post-war period. And yet, GDP grew by $1 trillion
during the first two years after the full tax cut.
GDP went up by over $1,100 per person, and taxes rose by
what today would be about $80 billion a year. So I think there
is evidence in these two bipartisan cases that dramatic action
on the deficit, or a revenue-neutral tax reform if it
substantially lowers rates and makes the system more efficient,
that in those two cases that we should strongly look at dynamic
scoring.
[The prepared statement of Dr. Gramm appears in the
Submissions for the Record on page 38.]
Chairman Coats. Senator Gramm, thank you.
Dr. Hassett.
STATEMENT OF DR. KEVIN A. HASSETT, DIRECTOR OF ECONOMIC POLICY
STUDIES, AMERICAN ENTERPRISE INSTITUTE, WASHINGTON, DC
Dr. Hassett. Thank you, Mr. Chairman, and Members of the
Committee.
My written testimony, which I guess is perhaps way too
long, discusses the likely scale of economic impact of a
significant tax reform, and gives actually some mathematical
analysis of how wrong a static score can be, and for a typical
capital income tax reform I discuss evidence that we could
expect the static score to be off by about a factor of two.
And so then the question is: Should we adopt dynamic
scoring as part of the budget and fiscal policy process? And in
the rest of my testimony I discuss those issues, and that is
what I will focus on in my oral remarks.
Dynamic scoring is not an unprecedented move for the
government. Many branches of government must make forecasts in
order to fulfill their statutory mandates. Even those forecasts
are by their nature uncertain.
The Federal Reserve, for instance, must formulate monetary
policy in the face of macro economic conditions that remain
uncertain in perpetuity, albeit to varying degrees. Its members
regularly document their own forecasts, and Federal Reserve
policy is set with an eye toward the impact that interest rate
changes would have on the economy.
The reliance of the Federal Reserve on economic models to
set monetary policy is not controversial, nor a partisan issue.
The absence of controversy regarding that reliance reveals a
logical problem facing those who would dispute the usefulness
of dynamic scoring for fiscal policy. For example, many tax
reforms influence the economy by changing the cost of capital,
a variable that depends on expected tax rates, depreciation
rates, inflation, and the interest rate.
The Fed tracks the economic impact of interest rate changes
in part through the model of the cost of capital which
influences business investments and other decisions.
An identical change in the cost of capital can be generated
either through a change in the interest rate or through a
change in tax rates. The argument that it is acceptable to
model the effects of an interest rate change in one quarter of
the government with such a model but not to model the effects
of the tax rate change in another corner of the government
strikes me as simply illogical.
Though the context of dynamic scoring and the context of
monetary policy certainly are very different, in both cases the
proper responses for the forecast incorporate a nonpartisan
staff's best judgment of what the economic analysis shows.
The uncertainty economists face when evaluating fiscal
policy is not greater than the uncertainty that they face when
evaluating monetary policy. If we use models for one
application, we can use models for both.
As does the Fed in its analysis of economic conditions, so
should the staff of the Joint Tax Committee and others tasked
with dynamic scoring proposals incorporate sensitivity
analysis, a range of perspectives, and the best thinking of the
academic community.
If there are many available models for a specific question,
the staff should evaluate the broad range of them and then come
to a considered judgment regarding the relative weights of the
different results.
It is worth underscoring that this leaves discretion in the
hands of the Joint Tax Committee staff, rather than any one
model, in recognition of the necessity of human judgment in
formulating views based on economic models, rather than giving
one specific model the final word.
The last part of my testimony highlights a recent
development that is at a pretty advanced stage at the American
Enterprise Institute where we've thought very long and hard
about what it will take for dynamic scoring to be widely
accepted and also fully transparent.
And we have set up something that we call the ``Open Source
Policy Center'' where we have developed two types of models.
One is the type of model that draws on individual income tax
data and scores tax proposals on a static basis in a way that
is very, very similar to the types of scores that you get from
the Joint Tax Committee now.
And second, in collaboration with the BYU Macroeconomics
and Computational Laboratory and professors from BYU and MTSU,
we have developed a dynamic model of the economy which we have
bridged to the static model, and we have got a number of
collaborators all around the country and data users that are
already using these models.
They are fully open source, and every single assumption
that anyone could make can be tested and sensitivity analysis
can be performed.
It is our view that as we move toward dynamic scoring that
what we need to do is think of ways that we can take the
academic community and connect them to the policy community in
a fully transparent way. And at AEI we have made a large
commitment over the last few years to develop a fully open
source model that can be accessed by Congressional staffers and
even members themselves. We have got a Web interface to make it
easy to use.
We hope that the OSPC, the Open Source Policy Center,
evinces a level of transparency and technical rigor that serves
as an example for how dynamic scoring should proceed going
forward.
I think that it can be extremely productive to help us
think about policy if we see what the Joint Tax Committee or
the CBO's judgment about what the dynamic score is, and we know
why they made the judgment that they did, and we can test our
own judgment against it.
Congress and the United States would benefit more generally
from dynamic scoring. Much work remains to be done in fleshing
out exactly how such a system of dynamic scoring is going to
work in practice, but the obstacles to transitioning to a world
where it is done are not insurmountable. In fact, there is no
reason to delay the beginning of the implementation.
Thank you, Senator.
[The prepared statement of Dr. Hassett appears in the
Submissions for the Record on page 40.]
Chairman Coats. Dr. Hassett, thank you.
Dr. Diamond.
STATEMENT OF DR. JOHN W. DIAMOND, EDWARD A. AND HERMENA HANCOCK
KELLY FELLOW IN PUBLIC FINANCE, BAKER INSTITUTE OF PUBLIC
POLICY, RICE UNIVERSITY, HOUSTON, TX
Dr. Diamond. Chairman Coats, Ranking Member Maloney, and
Members of the Committee, thank you for inviting me to present
my views on the importance of dynamic analysis and dynamic
scoring.
Let me begin by reviewing the most recent budget
projections. CBO projects that under the extended baseline by
2040 revenues will be 19.4 percent of GDP, as opposed to 17.4
percent over the last 40 years; while spending will increase to
25.3 percent of GDP, as opposed to 20.1 percent over the last
40 years.
Clearly we have a spending problem. This implies that in
2040 the deficit would be 5.9 percent of GDP, and the federal
debt would be 103 percent of GDP. But this projection is far
from certain.
Under CBO's alternative fiscal scenario, the federal debt
is projected to reach 175 percent of GDP by 2040. There is also
uncertainty regarding behavioral parameters and the underlying
economic variables used in the projections.
In addition, there is uncertainty related to the economic
effects of enacting new policies. For example, CBO estimates
that including the macro economic effects of higher marginal
tax rates, larger deficits, larger transfer payments, and
increased federal investment would increase the projected
deficits in 2040 from 5.9 to 6.6 percent.
Currently, the economic effects of enacting new policies is
not considered in the budget process, even though other
assumptions in use create more uncertainty.
Why should we use dynamic analysis? Let me propose an
example. Consider two proposals. The first raises $200 billion
in revenue by taxing capital gains and dividends and increases
tax expenditures by $200 billion by expanding child credits.
The second proposal would raise $200 billion in revenue by
reducing child tax credits and reduce revenues by $200 billion
by lowering the capital gains and dividend tax rates. The
conventional estimates would view these two proposals as
roughly equivalent. However, analyses by JCT, the Office of Tax
Analysis, the OECD and myself and Alan Viard, clearly show that
the first proposal would decrease economic growth and increase
the deficits, while the second would increase economic growth
and lead to deficit reduction.
It is important that we account for these differences in
the policymaking process. There are several important issues
regarding how to implement dynamic analysis to improve the
budget process.
While providing a dynamic score is important, the primary
goal of dynamic analysis should be to compare the macro
economic effects of various provisions. And while examining
every provision on its own would be impossible--we do not have
enough time nor the resources on the staff--there are times
when it makes sense to examine a single provision.
For example, JCT recently provided a dynamic analysis of
the effects of permanently extending 50 percent bonus
depreciation and found that it would increase GDP by 0.2
percent over the budget window.
Another interesting study may be a look at a temporary
extension of bonus depreciation and a comparison of those two
policies. We must analyze proposals not only with positive
effects, or ones that we expect to have positive effects, but
we also need to analyze proposals that we expect to have
negative economic effects.
Identifying harmful proposals is just as important as
identifying proposals that increase economic growth. Dynamic
analysis should also examine the effects of related provisions
separately for large policy reforms.
For the BRT I examined the Tax Reform Act of 2014. It would
have been very interesting to split that analysis into three
separate analyses: one of corporate tax reform; one of a move
to a territorial system; and one of the effects of the
individual income tax reforms in that legislation.
Finally, let me just say that we need to include the debt
service cost in both the short and long run, that those effects
must be considered when we're looking at analyses of tax and
spending proposals.
It is also important to note that the macro economic
aggregates are not the only information that we should provide
to policymakers. Some measure of welfare is also important, or
a measure of the changes in distributional effects.
Finally, public disclosure is imperative and as much
information as possible should be released to the public. At a
minimum, enough information should be released so that outside
entities could replicate the work.
While dynamic analysis will provide valuable information
about the relative economic effects of alternative policies, it
will not solve the fiscal crisis facing the United States.
Policymakers will still face many tough decisions in the years
ahead.
Thank you.
[The prepared statement of Dr. Diamond appears in the
Submissions for the Record on page 49.]
Chairman Coats. Thank you.
And, Mr. Buckley.
STATEMENT OF JOHN L. BUCKLEY, FORMER CHIEF OF STAFF TO THE
JOINT COMMITTEE ON TAXATION, WASHINGTON, DC
Mr. Buckley. Thank you, Mr. Chairman and Ranking Member
Maloney for the opportunity to participate in your hearing
today.
I understand that the decision to use dynamic scoring has
largely been made by the Congress, but I believe there are
still issues remaining with respect to its implementation and
the interpretation of its results.
First, I think that the current state of the art when it
comes to macro economic analysis of changes in federal fiscal
policy simply does not provide the level of certainty or
consistency that is required in an official budget score.
For example, the Joint Committee on Taxation and the
Congressional Budget Office do not have common approaches to
the issue of dynamic scoring. Since both use different models
and different sets of assumptions, the same proposal could
receive dramatically different budget scores depending on which
entity did the scoring.
Also, it does not provide the consistency that Professor
Diamond suggests is necessary to compare competing proposals
because the difference in the score may only reflect which
entity is scoring the proposal.
Second, there are some models that I think are simply
unacceptable for being used in dynamic scoring. Those models
are called ``Forward-Looking Models.''
They are also the models that typically produce the
greatest growth effect. Those models have as an underlying
assumption that the Congress will enact deficit reduction
legislation in the future necessary to solve the budget problem
that Professor Diamond so accurately described.
I simply believe you cannot have the official budget score
of a piece of legislation dependent on the assumption that
Congress will do in the future what it is unwilling to do
today--and that is, to enact major deficit reduction
legislation.
Also, you cannot have the Congressional staff making
predictions of what you might do in the future on deficit
reduction. So those types of models I think you just have to
set aside and not use for dynamic scoring.
Also, I think you need to understand that all of these
macro economic models are mathematical formulas. They do not
attempt to measure the impact of the tax policy on our actual
economy, or based on actual human behavior. That would be far
too complex to measure to reflect in a mathematical formula, no
matter how complicated the formula would be.
Therefore, they measure the impact against a hypothetical
economy constructed through assumptions that are often counter-
factual in the sense that they are contrary to observable
facts.
Also, the basic theory in the models assumes that increases
in labor supply or capital, the factors of production, will
automatically translate into greater economic growth.
I think you have to question that theory. For example, I
think a simple question is: What is the biggest economic
challenge faced by this country?
Is it lack of job opportunities that could support a
middle-class family?
Or is it the fact that we have too few people looking for
work?
Now I know what I think is the answer to that question. It
is lack of job opportunities. Yet the models assume it is the
lack of people looking for work which is the economic problem
we are facing. They solve the issue of unemployment, or
underemployment simply by assuming that it does not exist.
Finally, Mr. Chairman, during the period after 1980 and
before 2009, we essentially ran a series of experiments on the
theories underlying the dynamic scoring models.
With one exception, the rate increases enacted in 1993,
almost all major tax legislation was consistent with the
theories that underlie these economic models. Essentially,
there was a real-life experiment whether those theories were
correct. The results proved that they were not.
The large marginal rate reductions were supposed to
increase savings. The savings rate declined precipitously. The
1993 tax increases were supposed to reduce labor supply. Labor
supply grew after those tax increases and reached a record in
2000.
The 2001 rate reductions were supposed to increase labor
supply. Labor supply began to decline.
So there is a long history here that I think the Committee
has to take into account and should discuss the basic
underlying premises of these models.
Thank you, Mr. Chairman.
[The prepared statement of Mr. Buckley appears in the
Submissions for the Record on page 60.]
Chairman Coats. Well thank you to our witnesses. A number
of questions have been raised here that I look forward to
having a discussion on.
I deeply regret that I have a conflict with the
Intelligence Committee briefing us on the Iranian Agreement
that I need to go to. So I am going to turn over the gavel to
Vice Chairman Brady in just a moment.
I first want to recognize Chairwoman Maloney, the Ranking
Member on the House side. I was hoping to delve into the
relationship of monetary changes enacted by the Fed in terms of
how that might affect the dynamic scoring model.
As we know, that information in terms of what the Fed may
be thinking and doing would not be available to us in terms of
going forward. We can take some guesses, but I would hope
someone would get into that answer. I wish I could be here. I
apologize for having to do double duty here.
But let me ask Congresswoman Maloney to give her opening
statement, and then I will turn it over to Vice Chairman Brady
for his statement, and then we will go to the questions.
OPENING STATEMENT OF HON. CAROLYN B. MALONEY, RANKING MEMBER, A
U.S. REPRESENTATIVE FROM NEW YORK
Representative Maloney. Thank you so much, Mr. Chairman,
and thank all of the panelists. Last month this committee held
a hearing on so-called ``fair value'' accounting, a method of
calculating the impact of federal lending programs that will
make them appear more expensive.
Today the Committee turns its attention to dynamic scoring,
a method of analyzing and quantifying the budgetary impact of
tax cuts that will make them appear less expensive.
Both methods are very problematic and in both cases they
change the rules of the game so my Republican colleagues can
get the results that they want. Dynamic scoring has been
conservatives' Holy Grail for many years. This is because if
tax cuts appear to cost less, it will be easier for Congress to
pass more of them.
But there are serious problems with dynamic scoring. One
problem is that it provides results that are highly uncertain,
vary widely, and could be subject to manipulation. Let's take
the example of former House Ways and Means Chair Dave Camp's
tax reform legislation in the last Congress.
The JCT performed a dynamic analysis to see how much
additional revenue the tax plan could return to the Treasury.
And it is up on the screen right now. They used eight different
models, and they came up with eight different answers, varying
from $50 billion to $700 billion. The largest estimate was 14
times the size of the smallest estimate. And which estimate did
Chairman Camp highlight? $700 billion, the highest one. This
leads to two more serious problems with dynamic scoring.
There is no consensus on which dynamic scoring model is the
most appropriate. And the models rely on assumptions that are
sometimes wildly unrealistic, as Mr. Buckley pointed out in his
testimony.
For example, one dynamic scoring model assumes that if the
debt increases as a share of the economy future Congresses will
deal with the problem. The model assumes that in the future
there will be no unemployment. The fact is that with dynamic
scoring budget analysts will be forced to choose between deeply
flawed models.
Former CBO Director Rudolph Penner has said that, and I
quote, ``Dynamic scoring would force analysts to make more
judgment calls than they do today. Quality control would be
difficult, and that implies a high risk that ideological biases
will pollute the analysis.'' End quote.
There is yet another serious issue with dynamic scoring.
New rules require a single estimate. Until now, the Joint
Committee on Taxation and CBO have been required at the request
of the Chairman of the Ways and Means Committee to provide a
range of dynamic analysis estimates to reflect the different
models and assumption choices. But the new rule passed by
Congressional Republicans requires JCT and CBO to provide a
single revenue projection, and the estimate is official, not
advisory. The example of Dave Camp's bill shows that dynamic
estimates for major tax bills can differ by hundreds of
billions of dollars. If the Camp bill had become law and the
$700 billion figure proved wrong, deficits would explode.
Because the results are so unreliable, dynamic scoring will
compromise the accuracy and integrity of the federal budgeting
process. Former Federal Reserve Chairman Paul Volcker has said
simply, and I quote, ``I won't believe the numbers.'' End
quote.
And what happens if the markets come to doubt the integrity
of the scoring process? Former Federal Reserve Chairman Alan
Greenspan has said that, and I quote, ``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.'' End quote.
Republicans' decisions to use dynamic scoring, a highly
unrealistic and deeply flawed method, may by itself have
negative consequences that overwhelm whatever positive revenue
effects that could be gained by cutting taxes.
There is still another problem with dynamic scoring as
implemented by this Congress. It strongly biases policy towards
tax cuts. The new rule applies dynamic scoring only to tax
cuts, not to discretionary spending.
There is a broad consensus among mainstream economists that
investments in infrastructure, education, and research and
development can have a strong stimulative effect. But the new
rules do not apply to discretionary spending.
For this reason, these investments will seem very expensive
relative to tax cuts, and Congress will be more likely to cut
them. But does this mean that we should apply dynamic scoring
to discretionary spending proposals as well?
No. Because an accurate and impartial method of dynamic
scoring remains far beyond the reach of economists and budget
analysts. Until those models improve vastly, there is little
justification for using dynamic scoring on either tax bills or
spending bills.
The dynamic scoring rule serves only one purpose. It helps
Republicans reach their Holy Grail, rigging the rules so it is
easier for Congress to cut taxes.
Bruce Bartlett, former aide to President Reagan, put it
this way: Dynamic scoring is, and I quote, ``is not about
honest revenue estimating, it's about smoke and mirrors to
institutionalize ideology.'' End quote.
I look forward to our witnesses' testimony and answering of
our questions. Thank you, and I yield back.
[The prepared statement of Representative Maloney appears
in the Submissions for the Record on page 33.]
Vice Chairman Brady [presiding]. Thank you, Mrs. Maloney.
I want to thank each of the witnesses for being here today.
For the sake of time, I have an opening statement that is so
compelling it will bring you to tears, but . . .
[Laughter.]
For the sake of the hearing, I will just submit it for the
record.
[The prepared statement of Vice Chairman Brady appears in
the Submissions for the Record on page 32.]
And I will point out that I was involved for the past three
years in Former Chairman Camp's tax draft proposal.
Now the reason there are a number of scoring elements is
that the first five are the routine static scoring models that
Joint Tax uses today, and in fact which our Democrat friends
rely upon for all of their proposals and did so.
The last two are dynamic scoring models which were used to
score the economic impacts of the Comprehensive Immigration
Bill, which our Democrat friends as well held as gospel. What
Chairman Camp was seeking to do was really use the range of
models available to Joint Tax today, but the newer, more real-
life models to try to estimate Chairman Camp's tax proposal.
Senator Gramm, you have obviously a key background in
economics. You have sat where we've sat, a leader of key fiscal
solutions on budget, on taxes, on financial services. We hope
to move tax reform that is comprehensive, that creates lower
corporate taxes, a territorial system, lowers the disincentives
for work and savings and investing in the United States, and do
so we hope over the next two years or so.
Sitting in our seats, you know, what is the single greatest
benefit you believe policymakers can gain from dynamic scoring?
Just how valuable is it to both sides as we weigh major fiscal
issues such as tax reform?
Dr. Gramm. Well, Mr. Chairman, first of all I think it is
unfortunate that the whole dynamic scoring debate is so focused
just on tax cuts.
Dynamic scoring is about trying to take into account the
impact of economic policy adopted by Congress on the economy,
and on revenues, and on spending.
It is far more than just a debate about tax cuts. It is
interesting, because Kevin made the point very convincingly to
me that the Federal Reserve Bank relies on estimates every
single day.
All of their policies are based on dynamic scoring as to
what they are trying to achieve. So whatever Paul Volcker said,
or Alan Greenspan said--and I respect both of them--every day
they worked at the Federal Reserve Bank they were dealing with
dynamic scoring in trying to implement monetary policy.
So the idea that we've got all these economists at the
Fed--Kevin was there; they were better in those days----
[Laughter.]
And we have got all these economists that are scoring
monetary policy that are looking at its impact on interest
rates, and growth, and employment, all of which are estimates,
all of which are imperfect, but they use it every single day
and nobody says anything about it. Nobody seems to think it is
unreasonable, but yet the idea that the Congress would do it
when we are changing the policies of the country, I think that
is an unreasonable position to take.
I think there are two cases where the empirical evidence is
pretty overwhelming that government policies have in the past
had some predictable impact on the economy and on revenues. And
I think one of them is dramatic reductions in the federal
deficit through spending control.
And I don't have any doubt whatsoever that a dramatic
reform of entitlements that affected the long-term deficit
position of the country would create incentives for people to
invest because of enhanced confidence and to consume. So--and
if you look at the Clinton-Republican Congress compromise, the
five years that the program went into effect, it outperformed
the economic impact estimated on a static basis by $2.4
trillion in GDP, and $1.1 trillion in tax collection.
I also think that the other case where the evidence is
strong is on the bipartisan 1986 Tax Reform Act. We were way
into the recovery. The economy was getting weaker. The
Congressional Budget Office was projecting a decline in the
growth rate. That tax reform gave a second wind to the economy
and clearly its impact was positive and the country benefitted
from it.
I think in those two cases, both of them were bipartisan
efforts, that the case is pretty strong empirically for the use
of dynamic scoring.
Vice Chairman Brady. Thank you, Senator. I am going to wrap
up my time, except I want to, one, congratulate Dr. Hassett on
the Open Source Model. I think it is critical for those who
have ideas on how we become competitive and grow this economy
to have models, to be able to plug those ideas in to look at
what that impact could be.
A quick question for both you and Dr. Diamond. You know,
the biggest criticism is that dynamic scoring is simply not
ready. That somehow the technology, economic knowhow simply are
not there.
Senator Gramm made the point, you know, in the 1980s and
the 1990s, static scoring missed it by a mile. And so can you
point out to us what advances have occurred over the past two
decades that make this more accurate in real-life for us?
Dr. Hassett. Thank you very much, Mr. Brady.
And, you know, there have been, and there constantly are
advances in our ability to model. But I think it is very, very
important to emphasize that as policymakers what you need to do
at a moment in time is set policy based on what the best
knowledge that we have is. That if we had perfect knowledge,
then all economists could retire and it might be a perfect
world with no economists, but we are going to constantly be
learning things.
So the notion that we are going to learn more should not be
an obstacle to using the best analysis that we have. And I can
give an example. I actually agreed with a lot of your
statement, Mrs. Maloney, and you raised some very important
concerns.
But let's think about the eight different models. And I
unfortunately did not see your slide ahead of time, but it
looked like all of the estimates of the impact of the plan were
that--were positive.
And so the notion that we have eight different models with
lots of different assumptions that assume that this responds a
lot, and that responds a little, and the other does the
opposite, and no matter how you look at it if all the models
are saying that there is a positive effect on growth, that the
dynamic score is below the static score, then it just does not
make sense to me to say that zero is the right answer.
And I envision a world where people that I have the highest
regard for, like Tom Bartol or Doug Elmendorf, look at a vast
amount of output. They look at the broad range of knowledge
that we have, and they make a considered judgment about what
the best answer is.
I absolutely share your concern that if a partisan person
were to make that judgment, then people would stop trusting it.
But I think that that is already a problem, right? So I think
that we do trust what CBO and JCT do right now because they've
got such a strong track record of hiring nonpartisan staff.
I do not think that allowing them to use their own economic
expertise to improve their judgments is going to change that.
Vice Chairman Brady. Thank you.
Dr. Diamond.
Dr. Diamond. I am not sure there are really any
advancements. What I would point to is just a fundamental
misunderstanding of modeling by detractors of dynamic scoring,
and Mr. Buckley just cited two of the most fundamental. Let's
start with the forward looking assumption. Critics say, people
cannot be forward looking, but I think we are all forward
looking.
Do you think about your future? Do you think about what is
going to happen in the future? I think we all do. So when we
are modeling, we have a choice. Do we want a model that assumes
people are forward looking? Yes, we know they are going to make
mistakes. But on average, their mistakes will cancel out. Some
people will assume that wages will be higher, and some people
will assume they will be lower, but on average the forward
looking assumption implies that people do not make the same
mistake. And they do not make the same mistake over and over
and over.
The models Mr. Buckley would like to use, called myopic
models, not only assume that people make the same mistake every
year, year after year, forever; it assumes that everybody makes
exactly the same mistake.
That is the worst model to use in many circumstances. And
he claims that the reason we should not use the models is
because the assumptions of forward looking models are
unrealistic. But on the first day of Standard Principles of
Economics, I always teach the same thing: Class, what we are
going to learn is that simplifying assumptions are necessary,
but they are not realistic.
What we need is a model that has predictive power, not a
model that looks exactly like the real world. What we want is
to predict things accurately.
Let's think about the most standard economic model. It
assumes things such as perfect markets. Many buyers and sellers
on both sides. Firms that sell perfectly identical products. No
barriers to entry. These are unrealistic assumptions. Should we
throw that model out? That model is the simple model of supply
and demand. Basing arguments on assumptions you do not
understand is not a good way to choose dynamic scoring models.
Vice Chairman Brady. Thank you, Dr. Diamond.
Vice Chairman Maloney, former Chairman.
Representative Maloney. Thank you so much, and I thank all
of you for your comments. But I would like to ask Mr. Buckley,
I would like to read a statement to you by Former Federal
Reserve Chair Alan Greenspan, and I quote, ``We should be
especially cautious about adopting technical scoring procedures
that might be susceptible to overly optimistic assessments of
the budgetary consequences of fiscal actions.'' End quote.
So if we applied dynamic scoring to tax cuts, is there a
risk that we could overestimate the government revenues?
Mr. Buckley.
Mr. Buckley. I believe there is. But let me first take the
opportunity to respond a little bit to what Professor Diamond
said.
My main objection to forward looking models is not
assumptions of forward looking. It is that they require an
assumption by the modeler that the Congress will enact deficit
reduction legislation in the future.
They don't score the bill before you. They score the bill
before you assuming that you will take action that as of yet no
one has been willing to put forward.
But to answer your question, I think you should be
conservative in budget estimates for the same reason that
corporations are not permitted to take into account the
benefits of their investments when reporting to shareholders.
The temptation to be overly optimistic is a little too large.
Even with independent auditors, no matter how certain the
corporation is that its investment will be quite profitable, it
has to record that investment at cost and take into account the
benefits when they accrue. And I believe that is the
conservative path that federal budgeting should also follow.
Representative Maloney. Okay, Dr. Diamond, would you like
to respond to Mr. Buckley's observations?
Dr. Diamond. Yes, I would. The second argument is that the
models are unreliable because they do not include rising debt
levels that cause the economy to blow up. Let's start with one
model that JCT uses. The MEG model assumes that Congress is
going to do nothing and that you are going to let the U.S. turn
into Greece.
That model is making an assumption; however, it is not a
very likely assumption. Forward looking models assume that
there is not a problem like in some of the conventional
estimates.
But here's the key----
Representative Maloney. But if----
Dr. Diamond. If we use----
Representative Maloney. If I could respond really briefly
because I have other questions, Mr. Buckley's point was that
you are assuming that you are going to do deficit reduction,
which I have not seen since I have been here, and that there
will be no unemployment, when of course there is unemployment.
But I do have a question that I would like----
Dr. Diamond. But let me--the deficit----
Representative Maloney. May I ask a question about your
report?
Dr. Diamond. Sure.
Representative Maloney. Your report on Chairman Camp's tax
reform plan, and on page 14, in this report that you prepared
for the Business Roundtable on the Camp Tax Reform Plan, you
said something very important and something I think this
Committee should listen to very carefully and very closely, and
I would like your response to it.
You noted that results of any one model are, and I am
quoting from you, that the results from any one model are,
quote, ``at best suggestive.'' End quote.
And what is the risk of basing revenue estimates on models
that are at best suggestive? Is there a consensus among
economists about which of these ``at best suggestive'' models
to use?
Dr. Diamond. I still stand behind that statement. Dynamic
analysis is at best suggestive. We can't produce a single
number with perfect confidence. But, you know what,
conventional analysis is at best suggestive.
Those numbers, as I highlighted in my opening testimony,
are extremely uncertain. Let's go back to the idea that the
problem is not forward looking models, but instead that forward
looking models don't include an exploding deficit. Let me tell
you why I do not include exploding deficits in my models--if I
include an exploding deficit in the model, so that we let tax
rates go from 17.4 in the model to 19.4, and we let spending go
from 20.1 to 25.3, and we continue to let the model explode
into the out-years, do you know what the effect would be on the
estimates of the growth effects of a tax cut if the model
started with much higher tax rates?
Standard economic theory says they would be much larger. If
I start at a zero tax rate and I increased taxes by one dollar,
the welfare effects are relatively small. If I started at a $10
tax rate and increased taxes by $1, the growth effects could be
a hundred times larger because it's the square of the tax rate
that matters.
My assumption moderates the results. It does not produce
larger results. It is a moderating assumption, and I know I am
right.
Representative Maloney. Okay----
Dr. Diamond. As far as my comment that dynamic analysis is
``at best suggestive,'' all estimates are at best suggestive.
That is why they are called estimates.
Representative Maloney. So, Mr. Buckley, what are your
thoughts about using models that are, quote, ``at best
suggestive''?
Mr. Buckley. Well I think there is a real need for macro
economic analysis in the development of legislation, and I
would suggest both tax and spending legislation.
The Congress should be informed on the consequences of what
they do. However, the best numbers are the broad range. That is
what the Congressional Budget Office has said, that the best we
can produce are broad ranges of estimates and you can judge.
But again, I think that for official budget scoring
purposes you must score the bill before you and nothing else.
You should not score the bill before you and the assumption
that you will make major reductions in entitlement programs in
the future. I like to think I have some political experience,
as well. I doubt that many Members would want to endorse an
economic plan the success of which was dependent on identified
cuts in entitlement programs--otherwise known as Social
Security and Medicare.
Representative Maloney. And, Mr. Buckley, forward looking
models like Professor Diamond's are built on the core
assumption that future Congresses won't allow increases in the
deficit as a share of the economy. Is this a realistic
assumption? And what are the implications of this?
Mr. Buckley. Well whether it is realistic or not, it is a
tremendous breach from your current practices of scoring only
the legislation in front of you. And I believe you should
continue that.
The range of models results from Chairman Camp's bill that
$700 billion was the forward looking model, which assumed
entitlement cuts.
Representative Maloney. And, you know, how do these
assumptions affect the likely accuracy of the models? For
example, the unemployment one, that in the future there is,
quote, ``no unemployment,'' or in the future that they're going
to cut, you know, the entitlements which has not really
happened?
Mr. Buckley. You know, this is where I think the
credibility of the numbers are at risk. And if the credibility
is lost I think there are potential adverse consequences that
could dwarf whatever the difference is between static and
dynamic scoring.
Representative Maloney. My time has expired. Thank you.
Vice Chairman Brady. Thank you.
Senator Cassidy.
Senator Cassidy. Mr. Buckley, now Senator Gramm mentioned
three criteria by which dynamic scoring should be judged: macro
economic theory conforms--empirically it has previously worked;
and that changes would accrue within the budget window of note.
Now do you disagree that that would be a reasonable--and
the burden of proof is upon those who desire the dynamic
scoring to say that this is the case? Do you feel as if under
that kind of guidelines dynamic scoring would not be helpful?
Mr. Buckley. You know, I don't disagree with Senator
Gramm's outline. I might disagree with ``empirical evidence.''
As I stated in my oral testimony--during the period between
1980 and 2009 there were a whole variety of tax bills enacted
that were designed to increase savings and labor supply.
Under the standard economic theory reflected in these
models, the dramatic reduction in marginal tax rates that
occurred during that period of time and the broad expansion of
savings incentives should have resulted in an increase in the
individual savings rate. It did not.
The savings rate dropped precipitously from 1981 to 2007
before the recession. Also, the thought is that labor supply
responds to increases or decreases in tax rates, the 1993 tax
increase in marginal rates was followed by a slow increase in
labor force participation rates.
Senator Cassidy. Now that--now, again, I feel like I am
speaking in front of folks who have fought these battles
personally, so, Senator Gramm, you had mentioned that in
particular, I think 1993 actually did achieve some degree of
entitlement reform with decreased deficit, and that was one of
the preconditions that you labeled would lead to an expansion--
i.e., a justification for dynamic scoring.
Do I understand that correctly?
Dr. Gramm. (Off microphone.)
Senator Cassidy. Your microphone, please.
Vice Chairman Brady. If you could get that microphone----
Dr. Gramm. Maybe I will do that. I want to be heard.
[Laughter.]
The economic growth rate was soft in the first two Clinton
years. Positive, but soft. The dramatic change came with the
bipartisan budget agreement and a reduction in the capital
gains tax and the child tax credit. And it was dramatic. Even
Clinton's budget before the balanced budget agreement was
projecting $100- to $200 billion deficit until Jesus came back.
But what happened was that by actually taking action on a
bipartisan basis that was credible--and people keep talking
about, well, you can't score based on what Congress might do? I
never heard of anybody propose that you do that.
Anybody that would do that is a moron, because Congress
talks and doesn't act. But when Congress did act, when you had
an Administration and a Congress committed to a policy of
controlling spending, you had dramatic economic results.
And the boom of the Clinton years occurred after that
program was adopted. And just to go back and make one other
point. Dynamic scoring as we're calling it, which means using
the best information available which may not be very good but
it's the best available, is done everywhere except here.
It is used in the private sector. It is used at the Federal
Reserve Bank every single day. Everything they operate on is
dynamic scoring. How can it make sense for them to do it and
you not to do it?
And finally, I believe there are some cases where you can
make convincing arguments--and they're not all related to tax
cuts. I think part of why everybody's talking past each other
is that this subject has become a surrogate for tax changes.
But dynamic scoring is not just about taxes. It is about
spending. It is about policy. It is about regulation. And the
idea that we ought to just completely write it off because we
are not perfect at it just violates every principle we see in
the world around us.
Every day we do the best we can with what we have. It is
not perfect, and probably never will be.
Senator Cassidy. Okay, thank you all. I yield back.
Vice Chairman Brady. Thank you.
Senator Klobuchar.
Senator Klobuchar. Thank you very much, Mr. Chairman. Thank
you to all of you. Especially welcome back, Senator Gramm. I
never got to serve with you, but thank you for being here.
I was focused on this just from some of the things that we
are working on right now in the Senate. We are of course
working--we are debating the Long Term Transportation Bill, the
Drive Act. I have been a supporter of that nearly from the
beginning because it is a six-year bill with three years paid
for.
And I was wondering how you see a bill like that, which is
just set pay-fors. Dynamic scoring would change the way we
would look at that bill.
And then secondly, another thing that we've been looking
for in the long haul, which is paying for infrastructure. And
Congressman Delaney, who is a member of this Committee, has a
similar proposal to Senator Warner's is looking at long-term
tax reform with international tax reform, because we have a
bunch of money, as you know, a trillion set overseas and we
want to try to figure out a way to bring some of that back. And
one thought is to link it into infrastructure.
So those are two things we have been debating in the
Senate, two different ways to handle transportation, one moving
and one kind of sitting out there that a lot of people would
like to do.
And so I just wondered how dynamic scoring would change the
way you would interpret those two different proposals. I guess
I would start with you, Senator Gramm.
Dr. Gramm. Well first of all, I think dynamic scoring would
be looked at for every proposed change in legislation. And to
the degree that you had a transportation bill that dramatically
changed the quality of transportation in the country, I think
that you could make an argument that it would have a macro
economic effect.
I think whether it would have an effect within the time
period you're budgeting, you would have to look at. But the
whole purpose of the transportation bill is to strengthen the
economy, to expand the Gross Domestic Product. I think it is a
perfect case of something that we would look at.
And I think you would go back and try to, for example, look
at evidence during the Eisenhower era when we built the
interstate highway system. I think that there might be
empirical evidence out there that could be used.
Senator Klobuchar. Why don't we go to you, Mr. Buckley, and
then work our way back.
Mr. Buckley. I am in agreement with Senator Gramm. I think
there is empirical evidence that infrastructure spending would
be valuable, and I think that type of information should be
part of the legislative process.
However, the economic models that we are talking about
today assume that infrastructure spending has an investment
return half of what would be available if it were not done,
compared to what the private sector would do.
The CBO simply assumes that all government investments have
a rate of return equal to half the rate of return realized by a
private investment. So if you use these models, which I believe
substitute assumptions for analysis, you would find that
infrastructure spending is not a good idea, and it provides no
benefit for the economy--which I think is just
counterintuitive.
Senator Klobuchar. Thank you. Dr. Hassett, do you agree
with that?
Dr. Hassett. Oh, thank you, Senator. You know, I think that
you are correct that the argument in favor, that Senator Gramm
and I and Mr. Diamond are making in favor of making the best
evidence available, should apply to a wide range of things.
The infrastructure literature is one of the strongest
literatures, I think, where it is very clear that
infrastructure investment on average has a very significant
positive growth effect.
But, you know, that gets back to my last point, and I
certainly don't want to take all your time, but if you think
about, we had the question of what is ``conservative scoring,''
and I'm not talking about partisan conservative/liberal, I just
say what is conservative scoring, it is an example.
So if you are going to spend money on something where the
estimated rates of return are in the double digits in the
academic literature, then you ought to get rewarded for making
such a good choice based on everything that economists know.
Senator Klobuchar. And Mr. Buckley has a different view of
that----
Dr. Hassett. He basically said something that made no sense
to me, frankly, that because if you are not going to do any
analysis then how are we substituting assumptions for analysis
by doing dynamic scoring? The whole point about not allowing a
dynamic score is we just assume--let me give you an example--
Senator Klobuchar. Do you have evidence with infrastructure
of how it has been scored in the past, like international tax
reform? And then Congressman Brady is going to ask----
Dr. Hassett. I would be happy to correspond on this. I did
not prepare an infrastructure----
Senator Klobuchar. Okay----
Dr. Hassett [continuing]. But as a conservative estimate,
just to give an example, suppose that we were to increase the
corporate tax rate to 90 percent. It is not something that
anyone would propose, but suppose that we did.
Well if we just do a static score of that, then we will get
a lot of revenue. And that is not a conservative judgment,
right? So the conservative should be like what is the actual
revenue that we can expect to get? That is what conservative
budgeting is. And not allowing a dynamic score I think is not
conservative.
Senator Klobuchar. Well I am out of time, but I might
follow up with some of this in writing, if you guys could look
at how these--I have just mentioned two separate proposals
here. You know, one is the Drive Act, which is the pay-for
model that Senators McConnell and Inhofe and Senator Boxer
negotiated. And then the other one is more of an idea of using
the taxes, the money that is sitting over there that we want to
bring in. And of course we have not really defined how much of
it would go to infrastructure, but it is just another way of
paying for it.
All right. Thank you.
Vice Chairman Brady. Thank you, Senator. Without objection
I will place in the record a report by Doug Holtz-Eakin, a good
friend of the Committee, on dynamic scoring and infrastructure
spending, how it is used in evaluating policy proposals.
[The report titled ``Dynamic Scoring and Infrastructure
Spending'' appears in the Submissions for the Record on page
69.]
Vice Chairman Brady. With that, Mr. Paulsen is recognized.
Representative Paulsen. Thank you.
It just seems to make sense that we should be using all the
tools available as we make these important policy decisions
that affect the lives of millions of Americans every day.
Because we live in a very dynamic world where businesses
and individuals make decisions in part based on what takes
place here in Washington, lawmakers should have access to
information that takes into account the real-world impact of
these proposed policies on the people we serve.
Congress does not have a good track record predicting the
economic impact of its policies, because we have relied on
these computer models that are unreliable. Everywhere else, as
you mentioned, we are using the best information available. But
for some reason we are not using it here, because we think we
are in some alternative universe, and so we don't have to worry
about that.
Senator Gramm, I think you pointed out the historical
concept of demonstrating how fiscal policy changes have either
accelerated or decelerated real GDP growth over the last
several decades, and how the resulting changes in economic
growth have affected federal outlays, receipts, budget
deficits, et cetera, with revenue being up, GDP being up after
some of the changes in the 1980s for instance.
I was actually encouraged just a couple of years ago when
the Senate, under Democratic control then, took a vote to have
dynamic scoring used as a part of their tax reform modeling,
and now the House has put this formally into its rules. I
strongly believe that we need to fix the broken tax code with
comprehensive reform so it promotes investment, savings, and
hard work.
So let me just ask this, and maybe Senator Gramm, I will
just start with you. Because you have been here as a former
Member with a wide variety of background, what value does
incorporating this real-world impact into a scoring model have
for current lawmakers? And do you believe the use of an
economic model that includes real-world or dynamic impacts
could help grow consensus here in Washington around tough-to-
tackle issues like tax reform or entitlement reform?
Dr. Gramm. Well I think that we need to use the best tools
that are available. And when dynamic scoring, as we're calling
it, but using the feedback effect that policy changes have on
the economy and on the Federal Government's fiscal position,
that refusing to look at that simply guarantees that we are
going to have poor results.
And as I said in my opening statement, if you look at the
budgets of the United States and what the predictions were and
what has happened, the biggest errors always occur because of
changes in the economy. And they just swamp policy changes that
are scored on a static basis.
So I can't understand why we would not try to undertake
this. And you've got to undertake it for everything. Trying to
look at feedback effects on the economy is not about tax cuts.
It is not about changes in transportation. It is about all the
above.
Now you have got to meet criteria, it seems to me, to claim
the scoring. You've got to have a theory that makes sense.
You've got--it's got to have a feedback effect in the time
period you are budgeting, and you've got to provide some
empirical evidence.
But where you can do all three, to just simply say that
this makes no sense, I think again where these terms become
proxies for policies that people differ from, if somebody could
come up with an education reform program that honest-to-God
dramatically affected education in America, and did it quickly,
it would merit a huge dynamic scoring.
Now there is a big difference between talking about it and
doing it. But the point is, those are the kind of things we
ought to be looking at. And if somebody has got a good idea,
they ought to get credit for it in terms of what it is likely
to produce. I think that is the point that we are making.
And I don't see how you can be for dynamic scoring for
transportation and not for dynamic scoring in tax reform. I
mean, again it is obvious. You are just talking about what you
are for, not for the tools you ought to use in trying to
understand it.
Representative Paulsen. Is there a downside to having the
additional information that dynamic scoring can provide?
Dr. Gramm. Well look. You can always be wrong, and we're
almost certainly going to be wrong, but it seems to me in every
area of life, from the practice of medicine to drilling for
oil, to whatever, you operate with the best tools you've got
until you get better tools, but you learn from the process. And
I think that is what we need to undertake.
And I like the idea of a range of options. I like the idea
of giving outside people a chance to comment on it. I like the
idea of trying to form a consensus. But I don't think you can
begin: Well, I'm for dynamic scoring here because I am for this
policy, but I am against it here because I am against that
policy. It just does not make any sense.
Vice Chairman Brady. Thank you. Representative Beyer, you
are recognized.
Representative Beyer. Thank you, Vice Chairman Brady. Thank
all of you very much for coming to be with us.
Senator Gramm, it is wonderful to see you again, and thank
you for your humility as an economist. I am encouraged by--
although I am hearing first that there seems to really be a
consensus among the panel that if dynamic scoring makes sense
on the revenue side, that it also makes sense on the investment
side, at least for things that can be measured like
infrastructure investment.
I would also like to thank Dr. Diamond for his comment, ``I
know I'm right.'' It's the first time I have ever heard an
economist say something with such confidence. So, excellent.
[Laughter.]
Good work. In Senator Gramm's written statement, and I
think also you said here there were three conditions. You said,
first there must be a clear and established economic theory
suggesting a causative link between specific policy changes and
a substantial macro economic effect, et cetera.
And Peter Orszag, in this thing that was handed out, said,
on dynamic scoring, ``You're forced in the organization to pick
one true model, when economic science hasn't produced a single
model that works.''
So I got to study economics for four years as an
undergraduate and am completely confused. We weren't supposed
to have stagflation ever, and Japan had it for 10 years. No
growth and strong inflation.
Our $800 billion stimulus bill put together with
quantitative easing one, two, three, and four, was supposed to
give us inflation, and we have not seen it.
IMF and Europe imposed austerity on Greece to fix their
economy. It clearly has not worked.
Mr. Buckley, do you think that we have established, quote,
``a clear and established economic theory that gives us a basis
for dynamic scoring''?
Mr. Buckley. I don't believe that any member of the panel
would say that there is a single model that comes up with the
right, acceptable number. So the answer is: There's not.
Now one thing I think, at least in my mind, there is a
sharp distinction between providing more information for the
debate and affecting official scores. I think the more
information, the better. In that broad range of estimates, it
is probably the best you can do.
I think on transportation spending, there is real good
evidence that it provides benefits that are dramatic to our
economy. You cannot have a modern economy without a modern
transportation system.
I think that information should be part of the debate--but
the question is: Would you reduce the cost of a transportation
bill by those benefits?
I think that would be inconsistent with cost accounting. At
the end of the day when they announce the deficit, those dollar
expenditures will be recorded, not reduced by anticipated
benefits.
So I think you have to be kind of consistent in the way you
do it. Now don't interpret me as saying you shouldn't be
provided that information to justify this.
Representative Beyer. That is a great transition. Back to
Senator Gramm, both in your questions and your written
statement you say, quote, ``It is important to remember that
dynamic scoring is not a replacement for traditional static
scoring, but rather an enhancement of it.'' But the new rule
passed by the Congressional Republicans in the House, as least,
H.R. 1, requires the JTC and the CBO to provide a single
revenue projection.
I am sort of building on what Congressman Paulsen said.
That estimate is official, not advisory. Wouldn't you agree
that this new rule makes less information available to
policymakers rather than more?
Dr. Gramm. Well I can't imagine that they're not going to
provide the building blocks they use to try to come up with the
scoring. If I were doing it, I would want to set some broad
parameters. And then I would want to try to see to what degree
you might reach a consensus as to what the best estimate would
be.
So I might go about it that way. But do I believe we're
better off in trying to look at the feedback effect of our
policy on the economy and the government? I think we are better
off trying to do it. I don't claim it's going to be perfect, or
it's going to be a good estimate every time. But, you know, you
look back at even static estimates we've made, often they've
been very poor estimates.
I could give you examples that would go on and on about how
we projected something and then the economy just blew it away.
Representative Beyer. Thank you, Senator Gramm.
Mr. Chairman, I yield back.
Vice Chairman Brady. Thank you.
Representative Schweikert, you are recognized.
Representative Schweikert. Thank you, Mr. Chairman.
Dr. Hassett, if I came to you right now and said I want to
design the optimum dynamic scoring model in a modern society,
and you know, I'm a big fan of crowd sourcing information and
today we all walk around with super computers in our pocket.
The ability to grab lots and lots and lots and lots of data
sources and capture them--because my understanding is you're
the closest one right now to sort of doing that in a public,
open forum. How far can we take that?
Dr. Hassett. Thank you, Mr. Schweikert. It actually is a
very relevant point for the previous conversation, too, because
my belief is that there is no one correct model, that there are
lots of models with lots of different characteristics. Some
models allow for unemployment. Some of them don't. They assume
that we're always at full employment.
I think that, you know, Mr. Buckley said we shouldn't use
the models, we shouldn't put any weight on the models that
don't have unemployment. And, you know, the current CBO long-
run forecast assumes that we're at full employment a couple of
years from now. So it's a very standard thing.
But what we have to do is let a professional staff look at
all of the evidence and then make a considered judgment about
what the right answer is. And that is the way to do it. And so
I want to look at models like Mr. Diamond's model, which is a
model I worked in graduate school, a predecessor model of that.
But there are a lot of other models, too, and some of them have
Keynesian effects, and----
Representative Schweikert. But would you accept an open
source model where different data sets could be put in? You
know, if I had some data set from my region of the country----
Dr. Hassett. Exactly.
Mr. Schweikert [continuing]. I could plug it in and see the
effect?
Dr. Hassett. And what you have to do is look at basically
the information set and construct estimates of what are the
probabilities of the different things that might happen. And
the way you do that is you look at lots of different people's
approaches.
And so what we have tried to do is, one of the obstacles
for this is there are macro economists all over the world
developing macro economic models that will allow you to change
policy and see what happens. But they are not linked to the
things that we use to score because they don't have the micro
simulation model as the sort of first move.
And so what we have done is that we have automated the
bridge between the model that you get right now in the static
score and the things you need to actually get a macro economic
model to work so that people with macro economic models can
link them to what we're doing, and then hopefully in a year or
two we will have lots and lots of these models that we can look
at.
Representative Schweikert. Dr. Diamond, (a) is that the
future of how you would do it, but also how do you design a
model that reflects today compared to the data set we had a
month ago? I am fixated on the Atlanta Fed's GDPNow because of
its constant reacting to what happened that week of data.
I mean, how dynamic can you make the model? And can you
make it in a way where we are able to look at it today and
understand what it is doing to our policies?
Dr. Diamond. I think what Kevin is doing in open source
modeling is invaluable, and it is an idea that I've kicked
around and just never made it work, and I am really glad to
hear that someone is taking the lead. I think it is going to be
a brilliant advancement of modeling technology.
It may--I don't think the model you are explaining really
exists today. I mean, in some sense they do but the changes are
hard, and they take a lot of time.
Representative Schweikert. But my concern, where I was
trying to take this is what happens today when we get
information that says, hey, the decision we made six months
ago, or five months ago, isn't working? Should we as a Congress
also start to become much more dynamic in our policy? Instead
of saying: This is our policy for all of 2016, and if it
doesn't work, well be damned with it.
Dr. Diamond. Absolutely. We should all be like that. And I
think that when we get that evidence, we have to be willing to
change course. And that is why I think dynamic analysis is so
important, because it provides information about which course
you want to take.
Representative Schweikert. Well you could also start to
design policy. It's as if the data you're getting does this,
the law kicks in this, or takes this away, or adds this. So you
could also actually start to be much more disciplined and
creative in what we actually draft around here.
Dr. Diamond. That's amazing.
Representative Schweikert. Senator, okay, you were trying
to make the point of how we've heard some fairly blatant
partisan discussion on dynamic scoring, but I still remember
many of my friends on the left just being almost evangelical
about dynamic scoring when they were talking about the $831
billion stimulus bill, and the multiplier effects it was going
to have.
Don't we have lots of examples around us where we seem to
choose our poison?
Dr. Gramm. Well I think, and God knows I don't want to be
critical of the Congress, but what tends to happen----
Representative Schweikert. Oh, please do.
[Laughter.]
Dr. Gramm [continuing]. Is that people pick and choose
based on what they want, sort of to try to get the best
argument they can make for their position. And it is easy to
understand, and I'm sure that I have done it on many occasions.
The point is, however, that this ought to be something we
are looking at all the time. And in most cases a group of
totally nonpartisan experts, if such a thing exists, would
throw it out and say: Well, this just doesn't rise to the level
that you would ever want to make a projection based on it. It's
not big enough. It doesn't happen soon enough. There's not
enough empirical evidence.
But every once in awhile there will be a policy change that
is big enough. And when it does happen, it ought to either get
credit if it has a positive effect, or have cost attributed to
it if it has a negative effect.
And when you were saying about how, you know, if we had
this evidence the policy was not working we'd quit doing it,
actually most of the arguments would be it's not working
because we're not doing enough.
Representative Schweikert. And with that, Mr. Chairman,
thank you.
Vice Chairman Brady. Thank you. Representative Delaney is
recognized for what will certainly be a discussion about
infrastructure----
Representative Delaney. Well, no, I feel like that was
covered thoroughly. So I want to pivot to just a question.
Because, look it, to me there is no argument against dynamic
scoring. Right now the Congress has put itself in a position
where it cannot make any judgment decisions, right?
We assume that changes in revenues have no effect--or
changes in tax policy have no effect on behavior; we know they
do, sometimes dramatic, sometimes modest.
We also assume that government spending and investments
have no effect on economic activity, and we know they do.
Sometimes dramatic. Sometimes modest. So to me there is no
legitimate economic rational analytical argument against
dynamic scoring. We should be doing it.
But when I think about it from kind of a private sector
context, when a private business changes its revenues, or it
changes its pricing to hopefully encourage more revenues--which
is kind of the analogy to tax policy--or when a private
enterprise makes an investment because it thinks it will have a
decent return on its investment instead of modeling it at a
zero, which is basically what the government does, there's a
governance model in place where people generally have the best
intentions in terms of making good rational decisions. So a
board of directors looks at a proposal to make an investment,
or they look at a proposal to lower pricing, and they debate
whether it will have the intended effect. Sometimes they're
right. Sometimes they're wrong. But there's a good governance
process where these decisions are made on a rational basis.
The worry, obviously, with dynamic scoring, the only worry
I have is it will obviously be manipulated for ideological
benefit.
So do you have any thoughts as to what other things should
change from a governance perspective so that we could actually
feel comfortable doing what we obviously should do, which is to
get away from static scoring, which we know is wrong 100
percent of the time? It's always wrong, right, because there
are these behavioral changes. There are these economic effects.
And move to dynamic scoring which has a much better chance of
being accurate. It's not going to be 100 percent accurate, but
it has a much better chance. How we can do that in a way with
some kind of comfort that maybe we have a better governance?
Dr. Hassett.
Dr. Hassett. May I answer first, because--and thank you. I
absolutely share that concern. And the first thing is that I
think that a static score you would basically have the same
concern, right, that they could call it a static score. And yet
our scoring bodies are incredibly distinguished. I trust them,
and I think you trust them to do the static score to very high
professional standards.
But the second thing, and this is the thing that's a little
bit different from the current static scoring practice that
we're trying to sort of insert ourselves into, is that we just
need to see how they do it----
Representative Delaney. Right.
Dr. Hassett [continuing]. What they say, because, you know,
to try go back and figure out whether scores were correct or
not, on average it's almost impossible. It's a very, very
difficult thing.
But if we start being fully transparent, then we could
evaluate how we do. We thought it was going to be this much
revenue, it was that much revenue.
Representative Delaney. So like a budget--you know, in the
private sector you would normally have, when you're looking at
financial performance, you have budget and actual. And you're
actually looking at how your performance compared to what you
thought it would be.
You would recommend more of that kind of discipline?
Dr. Hassett. Yes.
Representative Delaney. You're right, because there is a
bit of a man-behind-the-curtain thing here, which it's not
clear how some of these scores are determined. That will be
less transparent. Let's face it, static scoring is easier than
dynamic scoring, right?
Dr. Hassett. Um-hmm.
Representative Delaney. So this will be a harder process I
think. And we do need much more transparency of how they make
the decisions, and actually how they're performing. Because if
they're doing a bad job, we should get new scorers in, right?
You know, just like in a company, if people do bad financial
modeling, you get new modelers in to hopefully do a better job.
Dr. Hassett. Can I even say one last thing, which I feel
strongly about, that I'm not sure I would chose dynamic scoring
that's not transparent for static scoring, because with dynamic
scoring you have a lot more wiggle room to do stuff. And if
there is somebody unethical doing that, then it is going to be
much harder to discipline them. So I think it is very important
for dynamic scoring to be done in a fully transparent way.
Representative Delaney. So you would make them disclose all
their assumptions.
Dr. Hassett. Yes, and the model.
Dr. Gramm. And all their data.
Representative Delaney. Yes, that went behind it. So what
we are talking here is dynamic scoring coupled with a much more
robust level of transparency. Any other governance changes you
might make, Senator Gramm?
Dr. Gramm. I think because of partisanship, because of the
difference in the sort of behavioral objectives in a private
entity that at least everybody is trying to be successful
versus a political entity where people have different
objectives, I think you've got to have a pretty high standard
that has to be met before you are going to employ the result of
your model.
I think there is a heavier burden of proof here. Sort of an
effort to sanitize it where there is enough of a consensus that
there----
Representative Delaney. Right.
Dr. Gramm [continuing]. That there is more than just a
partisan push here. And I think again this open model where you
could get input from anybody in the world who could send you
their views on it, and, you know, a lot of them would be
pretty--you wouldn't take seriously, but some of them might be
very serious. And I think that's a good idea.
Representative Delaney. Good. Good. Thank you, very much.
Vice Chairman Brady. Thank you.
Representative Grothman for the final question.
Representative Grothman. I hate to question this love fest
here, but I would like to respond. I personally believe in the
Laffer curve. I do believe as you cut tax rates it has to
affect behavior.
I am very, very skeptical of studies that show that
everything the government does is an investment and will pay
for itself. You know, we need more preschool, more kids going
to college even though people can't get jobs today. We need
more prevention programs, more infrastructure.
And when you combine the idea that tax cuts result in
increased revenue collections, and that all these new spending
programs result in--are actually investments which will more
than pay for themselves, it seems to me, well, it seems to be
pretty keynesian economics, almost making it part of the
statutes. You know, the idea that the bigger and bigger
deficits just keep paying for themselves, I think that's a
little bit scary.
Obviously I dislike the idea that bigger government leads
to more prosperity, more than I do the idea that lower tax
rates lead to prosperity. But just on the face of it, it seems
where we're headed is, let's in the next budget cut taxes and
spend more money on a variety of programs, and we're just going
to be running surpluses soon.
And I think that is kind of a scary thing. I would like you
guys to respond to that fear that I have.
Dr. Gramm. Well I think it can be a scary thing. I think
that everybody argues that their pet program is the magic
solution. And I think that is why you've got to have a very
high standard before you would accept to use dynamic scoring.
Now most of these arguments fall apart when you take a
close look at them, but I think that setting up a procedure to
evaluate them, where you have agreed in advance that unless the
evidence is pretty overwhelming you are not going to do the
dynamic scoring, I think that is the right way to go.
But I think caution on both sides of the aisle is the right
thing to do.
Dr. Hassett. Could I just add one logical thing? It's very
short. That if we spent the $100 billion burying $100 billion
in the ground, the classic textbook Keynesian policy, we would
get more GDP this year. But then we don't do it next year and
government spending is going down by the $100 billion, so we've
just located $100 billion this year, then there is an equal and
opposite effect tomorrow and GDP is going down because we've
got less government spending than we had this year, and so the
growth rate will be lower. So we could spend more this year,
but then we have to pay for it. And when we pay for it, there
will be a net cost.
And so if you look at the long run effect of Keynesian
policies in a budgetary manner, then you find a negative
cumulative effect because there's equal and opposite effects up
front but then a long-run cost of paying for it. So I don't
think that it would induce a lot--if we were to increase
government spending, it would produce a lot of Keynesianism.
What it might do, though, is make you spend things where we
have a lot of evidence that it is a positive, like building
wider bridges and things like that. There is a lot of evidence
that that is a very high rate of return place.
Representative Grothman. We would all be wealthier if we
had wider bridges? Do you really believe that?
Dr. Hassett. It depends on where you put them, but there
are a lot of bottlenecks in the D.C. area where people are
wasting a huge amount of time getting to work because you have
to get across the river if you're in Virginia.
So, yeah, I think that you could make Virginians wealthier
and their property values would go up if it was easier to get
into D.C.
Representative Grothman. Would you describe yourself--I
mean, I didn't know we had--I guess we're told on the thing
here that, you know, the guys on the left are Republican, or my
left. Would you describe yourself as a Keynesian?
Dr. Hassett. No. Absolutely not.
Representative Grothman. Okay, I will yield back my
remaining 40 seconds.
Vice Chairman Brady. Thank you.
I want to thank the panel for being here today. Let me
first submit for the record for Mrs. Maloney a report from the
Center on Budget and Policy related to budget and tax plans, an
outline on dynamic scoring.
[The report titled ``House `Dynamic Scoring' Rule Likely
Will Mean More Tax Cuts--Not More Information'' appears in the
Submissions for the Record on page 87.]
Vice Chairman Brady. You know, our goal was to have a
discussion about how do you create the most accurate and
complete assessment of the economic impact of policies.
What I seemed to hear today was that dynamic scoring does
not apply to everything, but where it does it should be applied
and considered; that the impact has to be big enough; the cause
and the evidence has to be accurate enough. And it is critical
that all these models be open both in the data and the models
and the assumptions for both parties to have confidence in the
range that it is arriving at.
So with that, let me thank all the panelists for being here
today, and I want to give a special shout-out to my former
Senior Senator from Texas who we work hard to try to follow in
your footsteps every day. Thanks so much for coming back to the
Senate today.
With that, the hearing is adjourned.
(Whereupon, at 3:38 p.m., Tuesday, July 28, 2015, the
hearing was adjourned.)
SUBMISSIONS FOR THE RECORD
Prepared Statement of Hon. Dan Coats, Chairman, Joint Economic
Committee
The committee will come to order.
I would like to welcome our witnesses, including my former
colleague and good friend Phil Gramm, who I am not used to seeing on
the other side of the dais! I thank all of our witnesses for being here
today to discuss the concept of ``dynamic scoring,'' a topic that has
been much debated since the House passed a rule earlier this year
requiring the Congressional Budget Office and the Joint Committee on
Taxation to use dynamic scoring when evaluating ``major legislation.''
The Joint Committee on Taxation and Congressional Budget Office
have long provided lawmakers with estimates of spending and revenue
changes that would occur should a bill become law. For decades,
however, these ``scores,'' as they are known, have largely ignored the
biggest driver of surpluses and deficits: economic growth.
That's because the current method of estimation--known as ``static
scoring''--does not reflect the reality that the economy can grow or
contract as a result of public policy. Most notably, it does not
account for the massive effects that policy can have on labor supply or
private investment, two of the largest drivers of the U.S. economy.
Ignoring these effects leaves lawmakers in the dark, unable to debate
legislation with all available information at our disposal.
While dynamic scoring has been debated for decades, it is no longer
``voodoo economics.'' In fact, advances in computer technology and
economics have finally brought us from the question of, ``Can it be
done?'' to the answer of, ``Yes, and here's how.''
We have the rare opportunity today to hear from those who have been
in the trenches of this debate as lawmakers, congressional staffers,
and academics.
I'd now like to recognize Ranking Member Maloney for her opening
statement and then will turn to Vice Chairman Brady, who was
instrumental in putting together this hearing.
__________
Prepared Statement of Hon. Kevin Brady, Vice Chairman, Joint Economic
Committee
Chairman Coats, Ranking Member Maloney, Members, and Distinguished
Witnesses:
Thank you, Chairman Coats, for convening a hearing on such an
important topic.
Let us begin with a common-sense proposition. To make fiscal policy
decisions that will increase the prosperity of the American people,
Congress needs to have the most accurate and complete assessment of the
economic effects of any proposed entitlement spending and tax
legislation.
Until this year, the Congressional Budget Office (CBO) and the
Joint Committee on Taxation (JCT) have ``scored'' proposed entitlement
spending and tax bills, respectively, on a micro-dynamic, but macro-
static basis. Under this treatment, the CBO and the JCT allow certain
changes in the economic behavior of individuals and businesses in
response to the enactment and implementation of proposed legislation,
but hold the size of the U.S. economy (real GDP) unchanged. For
example, the JCT would concede that a $5 per gallon increase in the
federal tax on motor vehicle fuels would cause households to drive less
and consume less gasoline. Counterintuitively, however, the JCT would
deny that such a tax increase would affect the U.S. economy overall.
This scoring convention is, of course, economic nonsense. In his
testimony Senator Phil Gramm demonstrates how major fiscal policy
changes accelerated or decelerated real GDP growth over the last three
decades and, in turn, how the resulting changes in economic growth
affected federal outlays, receipts, budget deficits (or surpluses), and
debt held by the public over the last several decades. Rather than
delivering realistic projections, the current scoring convention
reflects the limitations of economic modeling and computing capacity in
the 1970s.
In contrast with conventional scoring, dynamic scoring requires the
CBO and the JCT to assess not only whether proposed entitlement
spending and tax legislation would affect the economic behavior of
individuals and businesses at a micro level, but also whether the
aggregation of all such behavioral changes would affect overall
economic growth. In other words, dynamic scoring removes the
artificial, arbitrary, and unrealistic supposition that major
entitlement spending and tax changes will not affect the U.S. economy
as a whole. Put simply, dynamic scoring is proven, real-life analysis
that helps policymakers from both parties weigh the impact of proposed
changes.
The question before the Joint Economic Committee today is whether
the implementation of dynamic scoring of major entitlement spending and
tax bills would improve the quality of economic information available
to Congress before making major fiscal policy decisions. The answer is
an unqualified yes.
Since being elected to the House of Representatives in 1996, I have
been involved with tax and entitlement scoring issues on the Ways and
Means Committee under Chairmen Bill Thomas, David Camp, and Paul Ryan.
I have observed the great progress that economists from diverse
political viewpoints have made in refining their macroeconomic models
and developing a consensus around the estimates of key parameters over
the last two decades.
As Dr. John Diamond and Dr. Kevin Hassett will testify, economists
now have the ability to make reliable forecasts of the macroeconomic
effects of entitlement spending and tax bills on real GDP growth and
the feedback of such growth on federal outlays, receipts, budget
deficits (or surpluses), and debt held by the public. The limitations
that led to conventional scoring in the 1970s no longer apply.
Since 1997, the House of Representatives has allowed the Chairman
of the Ways and Means Committee to request dynamic analysis of major
tax legislation from the JCT, but for informational purposes only. In
2015, the House adopted a new rule requiring any proposed entitlement
spending or tax legislation that would create a gross budget change
equal to or more than \1/4\ of one percent of GDP to be scored on a
dynamic basis. Other legislation designated by either the Budget
Committee Chair or the Ways and Means Committee Chair must also be
scored on a dynamic basis.
Technology has advanced. The economy has become more complex.
Sticking blindly to the old ways robs policy makers in Congress of new,
more accurate insights on key challenges facing our country.
While dynamic scoring may involve multiple models and different
estimates of key parameters, dynamic scoring provides Congress with a
consistent, though not identical, view of how proposed entitlement and
tax changes would actually affect the real world. Yes, there is some
uncertainty, but that is part of the real world, too.
Currently, dynamic scoring applies to major entitlement reform and
tax legislation. One of the Members of this Committee, Representative
John Delaney, suggested in a Washington Post op-ed in January of this
year that dynamic scoring should also be applied to infrastructure
spending. While there may be merit to scoring government spending if it
significantly changes the overall economy, at this point Congress
should focus the CBO and the JCT on major tax and entitlement proposals
before expanding the scope of dynamic scoring.
I look forward to today's discussion with our witnesses.
__________
Preparted Statement of Hon. Carolyn B. Maloney, Ranking Democrat, Joint
Economic Committee
Last month, this committee held a hearing on so-called ``fair value
accounting,'' a method of calculating the impact of federal lending
programs that will make them appear more expensive.
Today, the committee turns its attention to dynamic scoring, a
method of analyzing and quantifying the budgetary impact of tax cuts
that will make them appear less expensive.
Both methods are very problematic.
And in both cases, they change the rules of the game so my
Republican colleagues can get the results they want.
Dynamic scoring has been conservatives' Holy Grail for many years.
This is because if tax cuts appear to cost less, it will be easier for
Congress to pass more of them.
Revenue estimates are based on projections of future behavior. For
many decades, budget effects from legislation were estimated using what
my Republican colleagues mistakenly called ``static'' models. These
models are not ``static'' because they anticipate how individuals would
react to the legislation, and the models are broadly-accepted by the
experts in the field.
Recently my Republican colleagues changed the scoring rule by
requiring the estimates to include the effect of legislations on the
whole economy, which is called ``dynamic scoring.''
But there are serious problems with dynamic scoring. One problem is
that it provides results that are highly uncertain, vary wildly, and
could be subject to manipulation.
Let's take the example of former House Ways and Means Chairman Dave
Camp's tax reform legislation last Congress.
The JCT performed a dynamic analysis to see how much additional
revenue the tax plan could return to the Treasury.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
They used eight different models. They came up with eight different
answers--from $50 billion to $700 billion. The largest estimate was 14
times the size of the smallest estimate.
Which estimate did Chairman Camp highlight?
$700 billion. The HIGHEST one.
This leads to two more serious problems with dynamic scoring--there
is no consensus on which dynamic scoring model is the most appropriate,
and the models rely on assumptions that are sometimes wildly
unrealistic.
For example, one dynamic scoring model assumes that if the debt
increases as a share of the economy future Congresses will deal with
the problem.
The model assumes that in the future there will be no unemployment.
The fact is that with dynamic scoring, budget analysts will be
forced to choose between deeply flawed, models.
Former CBO Director Rudolph Penner has said that:
`` . . . dynamic scoring would force analysts to make many more
judgment calls than they do today. Quality control would be
difficult, and that implies a high risk that ideological biases
will pollute the analysis.''
There is yet another serious issue with dynamic scoring--new rules
require a single estimate.
Until now, JCT and CBO have been required--at the request of the
chairman of the Ways and Means Committee--to provide a range of dynamic
analysis estimates to reflect the different models and assumptions
choices.
But the new rule passed by Congressional Republicans requires JCT
and CBO to provide a single revenue projection, and the estimate is
official, not advisory.
The example of Dave Camp's bill shows that dynamic estimates for
major tax bills can differ by hundreds of billions of dollars.
If the Camp bill had become law and the $700 billion figure proved
wrong, deficits would explode.
Because the results are so unreliable, dynamic scoring will
compromise the accuracy and integrity of the federal budgeting process.
Former Federal Reserve Chairman Paul Volcker has said simply:
``I won't believe the numbers.''
And what happens if the markets come to doubt the integrity of the
scoring process?
Former Federal Reserve Chairman Alan Greenspan has said that:
``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.''
Republicans' decision to use dynamic scoring--a highly unrealistic
and deeply flawed method--may by itself have negative consequences that
overwhelm whatever positive revenue effects that could be gained by
cutting taxes.
There is still another problem with dynamic scoring as implemented
by this Congress--it strongly biases policy toward tax cuts.
The new rule applies dynamic scoring only to tax cuts, not to
discretionary spending.
There is a broad consensus among mainstream economists that
investments in infrastructure, education, and research and development
can have a strong stimulative effect, but the new rules do not apply to
discretionary spending. For this reason, these investments will seem
very expensive relative to tax cuts, and Congress will be more likely
to cut them.
But does that mean that we should apply dynamic scoring to
discretionary spending proposals as well?
No--because an accurate and impartial method of dynamic scoring
remains far beyond the reach of economists and budget analysts.
Until those models improve vastly, there is little justification
for using dynamic scoring on either tax bills or spending bills.
The dynamic scoring rule serves only one purpose--it helps
Republicans reach their Holy Grail . . .
. . . rigging the rules so it's easier for Congress to cut taxes.
Bruce Bartlett, a former aide to President Reagan, put it this way:
dynamic scoring
`` . . . is not about honest revenue-estimating. It's about
using smoke and mirrors to institutionalize Republican
ideology.''
I look forward to our witnesses' testimony.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Prepared Statement of Senator Phil Gramm
It is a great honor to be asked to testify before the Joint
Economic Committee today, especially because I served with Chairman
Coats for many years in the Senate, and Vice Chairman Brady is an old
friend of mine from Texas.
During my time in the House and Senate, I focused mostly on the
economy and the budget. Anyone who spends any significant time studying
the U.S. budget comes to realize that changes in America's economic
performance have a profound impact on the budget of the country.
Economic changes often overwhelm the expected static impact of even the
largest policy changes.
Until we learn how to incorporate the impact of our policy changes
on the economy and the budget, we won't have a real understanding of
the costs and benefits of our proposed policy changes. When we have a
strong reason to believe that a policy change is likely to affect the
economy, based upon a logically consistent theory, and good empirical
evidence that similar policies have had significant effects on the
economy in the past, we should always attempt to employ dynamic
scoring.
Dynamic scoring is about finding a way to gauge the full impact
that policies might have in increasing or decreasing government
revenues and government expenditures. It seems to me that there are
three conditions that should be met before dynamic scoring can be used.
First, there must be a clear and established economic theory
suggesting a causative link between specific policy changes and a
substantial macroeconomic effect of sufficient magnitude to alter
revenues or outlays in the federal budget. Second, there should have to
be a good reason to believe that the macroeconomic effects would alter
government spending or revenues within the years that you are budgeting
for, which is normally 10 years or less. Third, there must be
convincing empirical evidence that the implementation of these policies
in the past has produced both the economic and the budgetary effects
that the theory would suggest. On all these points, the burden of proof
should fall on those who want to use dynamic scoring.
I'd like to discuss two compelling cases where the theory and
evidence of macroeconomic effects and budgetary feedbacks are strongly
supported. Both examples are bipartisan efforts and both relate
directly to topics that are at the center of the public policy debate
today.
The Balanced Budget Act and the Taxpayer Relief Act of 1997 was an
agreement between the Republican Congress and President Bill Clinton to
balance the budget through spending restraint while cutting taxes.
These bills had significant macroeconomic effects that benefited the
American people and the federal treasury alike.
In early 1995, CBO initially projected that balancing the budget by
constraining spending would create a combined revenue and outlay
dividend of $120 billion from 1995 to 2001, an estimate later increased
to $222 billion. After two years passed in negotiating the details of a
balanced budget deal, CBO reported in January of 1997 that much of the
original dividend had been incorporated into their baseline so that any
additional outlay and revenue dividend was just $43 billion for 1997 to
2001.
When we compare CBO's January 1997 GDP and revenue forecast prior
to enactment of the Balanced Budget Act to the actual results achieved
in the next five years, we find that both economic growth and revenue
growth after the Balanced Budget Act became law far outperformed
anything projected by CBO. Nominal GDP, from 1997 to 2001, surpassed
CBO's projected GDP by an astonishing total of $2.4 trillion--
equivalent to $4.7 trillion in today's economy (2014 GDP). That
averaged out to $480 billion per year higher than CBO's original
projections, providing an extra $8,609 in per capita GDP in those five
years.
Revenues also rose beyond expectations, even after Congress and the
President cut the capital gains tax rate and established the child tax
credit. From 1997 to 2001, cumulative federal revenues were $1.015
trillion higher than projected before the enactment of these laws. A
similar revenue surge today would deliver an additional $368 billion
per year to the government. The CBO reported in July 2000 that
``projected revenues for [FY] 2000 are now $303 billion more than
estimated in 1997 . . . The primary contributors to that unexpected
growth stems from the strength of the economy and changes in the
characteristics of income.''
The Tax Reform Act of 1986 was designed to be revenue neutral under
static scoring by closing loopholes and limiting deductions in exchange
for lowering tax rates from a top rate of 50 percent in 1986 to 28
percent starting in 1988. In comparison to CBO economic and revenue
projections prior to the full marginal rate reductions, the Tax Reform
Act produced a significant macroeconomic and budgetary impact. Its
benefits are magnified by the fact that this occurred well into one of
the strongest and longest postwar recoveries. By January 1988, the
recovery was in its 62nd month, over a third longer than the average
postwar recovery's length, with the economy averaging a scorching 4.6
percent growth and never less than 3.5 percent in any year.
Just prior to full implementation of the rate reductions, CBO's
economic projections assumed much lower growth, with estimated real GDP
growth of 2.3 percent and 2.6 percent, respectively, for 1988 and 1989,
but actual growth rates hit 3.9 percent and 3 percent (subsequently
revised to 4.2 percent and 3.7 percent).
Nominal GDP for those years surpassed CBO's projected GDP by a
total of $286 billion, equivalent to $1 trillion in today's economy
(2014 GDP). By averaging $143 billion per year higher, that benefited
every man woman and child in America on average by an extra $1,163 in
GDP during those two years. The Tax Reform Act gave a very strong
second wind to the recovery, helping to deliver a 38 percent increase
in real GDP in the 1982-90 recovery.
The stronger economy fed back into stronger revenues with Federal
income in the first two years after the marginal rate reductions
averaging $25 billion higher than expected. CBO reported that these
higher revenues were due to stronger economic factors. As a share of
2014 revenues, that $25 billion corresponds to $80 billion today.
Based on the evidence of the bipartisan Balanced Budget Act of
1997, we could expect that any dramatic change in budget policy that
substantially reduces the long term deficit through spending control,
such as spending restraint and entitlement reform, could reasonably be
expected to deliver substantial macroeconomic effects coming from
improved business and consumer confidence. I believe a very strong case
can be made that a comprehensive entitlement reform package that
dramatically reduced the long-term deficit should receive a large
positive dynamic score.
Similarly, based on our experience with the bipartisan Tax Reform
Act of 1986, we should have confidence in believing that revenue-
neutral tax reform that makes our tax system more economically
efficient and lowers tax rates would have a substantially positive
effect on GDP and, therefore, federal revenues. This is especially true
today given that the recovery of 2009 has never taken off.
The Joint Committee on Taxation (JCT) has already projected a
potential dynamic score of up to $700 billion over 10 years from one
version of pro-growth tax reform, which would correspond to an average
annual revenue increase of $70 billion. With a dramatic tax
simplification and rate reduction program, we could expect to achieve
dramatically positive results.
It is important to remember that dynamic scoring is not a
replacement for traditional static scoring, but rather an enhancement
of it. CBO and JCT have decades of experience estimating the direct
impact of legislative changes on the budget, but the largest revisions
to their projections and final figures have come from a failure to
fully predict and incorporate macroeconomic effects in their estimates.
Yet it is those very macroeconomic effects that have been so powerful
as to swamp the static estimates of the largest legislative changes.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Questions for the Record for Dr. John Diamond from Representative
Carolyn B. Maloney, Ranking Member
1. The Joint Committee on Taxation (JCT) used 8 different dynamic
scoring models to estimate the possible revenue effects of Rep. Camp's
tax reform proposals. These models predict that the proposal could
generate $50, $100, $150, $200, $225, $275, $650 or $700 billion of
additional revenue from macroeconomic feedbacks over 10 years.
It has been suggested at the hearing that five of these are
``routine static scoring'' models. Is that correct?
To begin with, JCT refers to its normal scoring method as a
``conventional'' revenue estimate not a static estimate. Given that, no
it is not correct.
JCX-22-14 states the following:
``The proposal is projected to result in increases in economic
activity relative to that projected under present law, as
measured by changes in real GDP. The increase in projected
economic activity is projected to increase revenues relative to
the conventional revenue estimate by $50 to $700 billion,
depending on which modeling assumptions are used, over the 10-
year budget period.''
Thus, all of the estimates are from macroeconomic analysis of TRA
2014. These estimates suggest that a conventional revenue estimate
would be wrong by a minimum of $50 billion and by as much as $700
billion. Taking one of the middle estimates of $200 to $225 billion
would be a reasonable number to use for dynamic scoring purposes.
Diamond and Zodrow (2014) also examined the effects of a proposal
similar to TRA 2014, which was prepared for the Business Round Table.
They found that the most important factor is the reduction in income
shifting as the corporate income tax rate declines. In addition, other
important factors include the move to territorial, the more efficient
allocation of capital, and the reallocation of FSK. Diamond and Zodrow
found that a proposal similar in structure to IRA 2014 would increase
GDP by 1.2 percent after five years, by 2.2 percent after 10 years, and
by 3.1 percent in the long run. The long-run increase in GDP is
primarily driven by a 5.0 percent increase in the ordinary capital
stock and a 0.3 percent increase labor supply. In the long run, a 57
percent reduction in income shifting allows the corporate income tax
rate to decline an extra 5 percentage points (relative to the Camp
proposal) to 19.9 percent (which is the only difference in the proposal
simulated and TRA 2014).
2. Dynamic scoring yields widely different results depending on the
model that is used and the assumptions on which the calculations are
based. During the hearing, you conceded that ``We can't produce a
single number.''
This problem is especially clear in the case of the Camp proposal--
the results of JCT models ranged from $50 billion to $700 billion. The
larger estimate--the one Camp cited--is 14 times larger than the
smallest one.
Nevertheless, Congress has passed a rule that requires JCT to
provide a single result. Policymakers will rely on this figure to make
decisions that could have enormous impact on our economy and our
budget.
If you believe ``we can't produce a single number,'' what are the
justifications for including these highly uncertain results in the
official score as a single number, rather than as a range of estimates
for advisory purpose in the conventional approach?
I firmly believe that we can produce a single number, however this
should not be the primary or sole goal of dynamic analysis. My
statement in the hearing (``we can't produce a single number'') was
simply a misstatement, which I have asked to be corrected for the final
record. My belief is that we should not only produce a single estimate,
but should examine a range of estimates to inform policymakers before
arriving at a dynamic score (a single estimate). Note that conventional
estimation also requires modelers to make assumptions and to pick a
single number from a range of possible estimates. As I stated in my
spoken and written testimony, there is substantial uncertainty in
conventional estimates. CBO routinely shows the significant uncertainty
that is contained in conventional estimates from policy uncertainty,
economic uncertainty, parameter uncertainty, as well as uncertainty
related to the economic effects of policy enactment. For TRA 2014,
simply assuming that there is no economic effect of the policy, even
though the estimated range is from $50 to $700 billion, is also
precisely wrong. Note that zero is a single number estimate!
The justification for including dynamic effects is to more
accurately account for the cost of various proposals. This is
especially important for policies that have negative economic effects,
that are debt financed, or that create large distortions. In addition,
dynamic analysis could play a critical role because budget gimmicks
within the budget window can often obscure the long-run effects of
policies, especially policies that are debt-financed, temporary, or
delayed and introduced late in the budget window. Ignoring the dynamic
effects would continue to allow the magnitude of the revenue effects of
many policies to be overstated and thus likely to lead to additional
debt. Note that huge deficits have occurred under a budget framework
based on conventional estimates, at least in part because such a
framework does not account for the true cost of competing policies.
For example, consider two hypothetical proposals. The first
proposal raises $200 billion in revenue by taxing capital gains and
dividends and increases tax expenditures by $200 billion by expanding
child tax credits. The second proposal would raise $200 billion by
reducing child tax credits and reduce revenues by $200 billion by
lowering capital gains and dividend tax rates. The conventional
estimates would view these two proposals as equivalent from a budget
perspective (with some small differences showing up to account for
certain timing effects). However, the above analyses clearly show that
the first proposal would decrease economic growth and cause an increase
in deficits, while the second would increase economic growth and lead
to deficit reduction if no other policy actions were taken. It is
important that we account for such differences in the policymaking
process.
3. During the hearing, Senator Phil Gramm dismissed Mr. Buckley's
assertion that one particular dynamic scoring model would force CBO and
JCT to predict what a future Congress might do. Senator Gramm said that
``Anybody that would do that is a moron, because Congress talks and
doesn't act.''
Do you see any problems in using a dynamic scoring model that
assumes how and when a future Congress will act to finance the deficit?
As I stated in the hearing, this is an important but extremely
misunderstood issue. It is widely recognized that under the current law
baseline U.S. fiscal policy is unsustainable. Let's start by thinking
about the effects of including various assumptions in dynamic models.
Within the budget window (the next 10 years), the effects are not
important. Current CBO projections show the debt to GDP ratio
increasing from 74 to 78 percent over the next 10 years. Failing to
account for such a difference in the baseline would not have a
discernable impact on the economic effects of enacting various policy
alternatives. The modelers can test this by running the policy change
at both debt levels. In the end, modelers can choose to use an average
level of debt in the model (e.g., 76 percent). In any case, the
differences of such assumptions would almost certainly be no more than
a rounding error in most cases.
In the long run, this assumption is more important. If you use a
model that allows for the enactment of an unsustainable fiscal policy,
then you must be keenly aware of when the model starts to diverge and
eventually it will fail to solve mathematically. During the divergence
process, the model will likely produce spurious results. This is
important to keep in mind when using such a model. In this case,
assuming current law (note that many commentators argue this is NOT the
most likely policy outcome, which is one reason CBO started producing
an estimate of current policy) requires the modeler to model turning
points in the economy (movements from booms to busts and busts to booms
are much more uncertain than predicting trend level growth) and the
effects of economic default. This is extremely difficult and would add
needless uncertainty to the process.
I do not see a significant issue with using a model that does not
allow for unsustainable fiscal policies in the baseline. In fact, I
believe this is a major advantage in one respect because it allows
modelers to ignore the future actions of policymakers, which are likely
unknowable, without having to model the effects of defaulting on
government debt. More importantly, in terms of measuring the economic
effects of policy changes, assuming away these large uncertainties
leads to more moderate projections (note that most critics, and
especially Mr. Buckley, are wildly confused on this issue). A basic
tenant of economics is that the excess burden or deadweight loss of a
tax increases with the square of the tax rate (that is, distortions
increase exponentially with increases in the tax rate). So the positive
economic effects from a tax rate reduction from an initially high tax
rate (such as those predicted in the U.S. under the unsustainable
current law baseline) would be much larger than a tax rate reduction
from a lower tax rate (those used in models that ignore the
unsustainable nature of current law). So assuming a baseline that
assumes taxes and spending are maintained at current levels as a
percentage of GDP are likely to moderate the predicted effects of tax
policy changes. It is my expert opinion that this is a desirable
feature of an economic model used to predict the effects of policy
changes well into the future.
There is another issue concerning the choice of fiscal offsets for
policies that are being examined. This is an important issue, however
the professional and knowledgeable staffs of the relevant committees
are more than capable of managing these issues.
Do you see any problems in using a dynamic scoring model that
assumes that every American who wants a job will be able to find one?
Absolutely not in the long run, and I do not think it is necessary
to worry about this problem in the short run as policymakers are
already incentivized to think about these issues in the short run. In
general, prices adjust to clear markets in the long run. If there is
excess demand for labor, then wages would be bid up. As wages rose
labor demanded would fall and labor supplied would rise. If there is
excess supply, then wages would be bid down. As wages fall labor
demanded would increase and labor supplied would decrease. While there
can be periods of disequilibrium (such as the period after the 2008
financial crisis), I am confident policymakers will continue to respond
with various short-term stabilizing policies during such periods (i.e.,
policymakers are often overly incentivized to focus on short run
issues). But such increases in unemployment are temporary and thus
should not be the focus of trying to determine the long run effects of
policies. This is a widely held view. For example, CBO (2014, p. 110)
states in the 2014 Long Term Budget Outlook that
``In its economic benchmark, CBO projects that real gross
domestic product will grow fairly quickly over the next few
years, reflecting a recovery in aggregate demand. Thereafter,
real GDP is projected to grow at a pace that reflects increases
in the capital stock, productivity, and the supply of labor.''
Thus, this states that in the long run CBO assumes that everyone
that wants a job will find a job. In addition, note that the Board of
Governors of the Federal Reserve is contemplating an interest rate
increase this fall at least partly because they see the labor market
returning to the full employment level (note that even in a fully
employed labor market there is still unemployment as some unemployment
is ``frictional'').
I do not believe long-term tax and spending policies should be
based on short run fluctuations in unemployment. Dynamic analysis
should be geared to adopting policies that maximize long run economic
growth. Thus, full employment models can and should be used to examine
the effects of tax and spending policies that are being adopted to
encourage long-run economic growth. There may be times when using
short-run models capable of examining the economic effects of policy
changes in markets that are not in equilibrium may be important, and at
that time those models should be used. But I don't believe we should
adopt long run policy on short run considerations.
4. At the hearing, you argued that ``What we need is a model that
has predictive power, not a model that looks like the real world. What
we want is to predict things accurately.''
What empirical evidence could you provide to substantiate your
claim that these dynamic scoring models ``predict things accurately?''
Can you please provide a list of peer-reviewed articles in reputable
journals that prove that these dynamic scoring models accurately
predict the revenue effects of tax cuts?
These models have been widely used and accepted by many in the
private sector, academics, and government. A great starting place to
answer your question is the Handbook of Computable General Equilibrium
Modeling, edited by Peter B. Dixon and Dale W. Jorgenson, published by
Elsevier (the most prestigious handbook series). Of course, my favorite
chapter is Chapter 11 by Diamond and Zodrow. The entire volume is a
great starting place to learn about the value of computable general
equilibrium modeling. Also, as stated in the hearing, economic models
are widely used across a wide range of government and private
institutions--it is well overdue for Congress to start using these
resources to make better policy decisions.
__________
Question for the Record for Mr. Buckley from Representative Carolyn B.
Maloney, Ranking Member
1. Given your analysis of the flaws of dynamic scoring models and
also given the fact that Congress has already adopted rules that
require a single dynamic score of large (tax) bills, would you
recommend Members of Congress to advocate dynamic scoring of both tax
cuts and discretionary spending? Are there any assumptions in the
dynamic scoring models that would bias against spending bills? Are
there any other potential pitfalls that policymakers should be aware
of?
If dynamic scoring is appropriate for large tax bills, I believe
that it is also appropriate for scoring major legislation that involves
spending on investments, such as infrastructure, research and
development, and education. Those investments provide substantial
benefits for our economy, benefits perhaps more certain than the often
predicted, but never realized, supply-side benefits of marginal rate
reductions. Without adequate infrastructure or an educated workforce,
the United States will not remain competitive in the world economy.
Also, if the scoring rules differ, there will be the temptation to
move spending programs into the tax laws where dynamic scoring would be
available. I assure you that spending through tax legislation is not
difficult to accomplish.
Unfortunately, current dynamic scoring models substitute
assumptions for analysis, and those assumptions are biased against
Federal spending, even needed spending such as investments on
transportation and other infrastructure.
For example, CBO assumes that the return on public investments will
be 50 percent of the return on private investments. They cite no
evidence for that assumption, but merely note that assumption is
halfway between zero used by some modelers and 100 percent used by some
others. Under their assumption, increased spending on public
infrastructure could be seen as negative for economic growth, something
that we know is not true.
[all]