Order Code RL31775
CRS Report for Congress
Received through the CRS Web
Do Budget Deficits Push Up Interest Rates
and Is This the Relevant Question?
Updated February 4, 2005
Marc Labonte
Analyst in Macroeconomics
Government and Finance Division
Congressional Research Service ˜ The Library of Congress
Do Budget Deficits Push Up Interest Rates
and Is This the Relevant Question?
Summary
With mounting budget deficits, attention has focused on their economic effect,
particularly whether budget deficits raise interest rates. Any explanation of the
budget deficit-interest rate relationship must first come to grips with an indisputable
fact: budget deficits consume real resources, and this — rather than the behavior of
interest rates — is the more relevant public policy concern. When the government
borrows from the public to finance public spending or tax cuts, the resources must
come from somewhere. In mainstream theory, the resources come from the nation’s
pool of saving, which pushes up interest rates for simple supply and demand reasons.
This “crowds out” private investment that was competing with government
borrowing for the same pool of national saving. For this reason, economists often
describe deficits as placing a burden on future generations.
But other theories offer different explanations of where the resources come from
that do not involve higher interest rates. In the capital mobility view, foreigners lend
the United States the savings it needs to finance a deficit, leaving interest rates
unaffected. But as foreign capital comes to the country, the dollar must appreciate.
This causes U.S. exports and import-competing industries to become less competitive
and the trade deficit to expand. In an alternative theory, popularly known as the
Barro-Ricardo view, forward-looking, rational, infinitely-lived individuals see that
a budget deficit would result in higher taxes or lower government spending in the
future. Therefore, they reduce their consumption and save more today. This
provides the government with the saving needed to finance its deficit, placing no
upward pressure on interest rates. Empirical evidence that budget deficits do not
affect interest rates does not prove that government budget deficits do not impose a
burden, as demonstrated by the capital mobility and Barro-Ricardo views. In the
capital mobility view, deficits crowd out the trade sector of the economy; in the
Barro-Ricardo view, they crowd out current private consumption. And in both of
these views, deficits no longer have any stimulative effect on the economy.
Comparing changes in budget deficits to changes in interest rates is not a valid
way to determine whether budget deficits affect interest rates. That is because there
are many other factors that also affect interest rates. To determine the effect of
budget deficits on interest rates, one must hold these other factors constant using
statistical methods. Otherwise, the effect of budget deficits on interest rates could
be misestimated or even reversed.
Empirical evidence on a link between budget deficits and interest rates is mixed.
There is not a consensus among economists on how to model the economy and what
relevant variables should be included. Therefore, conclusions drawn from empirical
evidence vary widely. More recent evidence tends to find a stronger, positive
relationship between the two. In addition, 10 major forecasting models all predict
that a budget deficit would increase interest rates. According to Gale and Orszag
(2002), the models predict that a budget deficit equal to 1% of GDP would increase
interest rates, with a range of 0.1-1 (mean=0.52) percentage points after one year and
0.05-2 (mean=0.99) percentage points after 10 years. This report will not be updated.
Contents
Budget Deficits in the Mainstream View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
The Conventional View With an Underemployed Economy . . . . . . . . . . . . . 4
Budget Deficits in The Mainstream View With Capital Mobility . . . . . . . . . . . . . 5
The Response of Saving to Budget Deficits and the Barro-Ricardo View . . . . . . 6
Saving in the Conventional View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Saving in the Barro-Ricardo View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Empirical Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Do Budget Deficits Push Up Interest Rates
and Is This the Relevant Question?
Budget deficits have been a concern of policymakers in recent years. In
FY2004, the federal budget deficit reached $412 billion. In the Congressional
Budget Office’s (CBO’s) baseline projection, the deficit under current policy (as of
January) is projected to be $368 billion in FY2005.1 This does not include the costs
associated with the war with Iraq, which CBO estimates might add another $30
billion in FY2005.
In response to these developments, policymakers’ attention has focused on what
effects these deficits are having on the U.S. economy. In particular, attention has
been focused on whether or not budget deficits raise interest rates. Some claim that
the deficits do indeed raise interest rates, to the economy’s detriment, while
advocates of recent tax cuts claim that there is no such link. This is an important
question because how effective tax cuts are in stimulating the economy depends on
how the rest of the economy adjusts.2 As this report explains, the adjustment can
occur through interest rates, the trade balance, or private saving. This report reviews
theoretical and empirical evidence to see what relationship exists between budget
deficits and interest rates.
Unfortunately, simply observing whether interest rates are higher or lower in the
presence of deficits is not a sufficient test of the hypothesis. That is because there are
many other factors that are simultaneously affecting interest rates. These include the
state of the economy, the saving behavior of individuals and companies, the
investment opportunities of business, the state of financial markets, demographics,
and the financial relationship between the United States and the rest of the world. In
the real world, these factors cannot be held constant, so economists try to “control”
for them statistically to separate out the effects of deficits. But controlling for other
factors requires a theoretical model that explains and identifies how a deficit and
those other factors affect interest rates. Thus, even those who are skeptical of
theoretical explanations are dependent on theoretical models to translate the
relationship from the empirical evidence. Because so many different models of the
economy exist, some models offer positive evidence about the deficit-interest rate
relationship while others offer evidence refuting the relationship.
Any explanation of the budget deficit-interest rate relationship must first come
to grips with an indisputable fact: budget deficits consume real resources. When the
government borrows from the public to finance public spending or tax cuts, the
1 U.S. Congressional Budget Office, Budget and Economic Outlook, Jan. 2005.
2 For evidence on the tax cuts’ efficacy, see CRS Report RL32502, What Effects Have the
Recent Tax Cuts Had on the Economy?, by Marc Labonte.
CRS-2
resources must come from somewhere. In mainstream theory, the resources come
from the nation’s pool of saving, which pushes up interest rates for simple supply
and demand reasons. This “crowds out” private investment that was competing with
government borrowing for the same pool of national saving. But other theories offer
different explanations of where the resources come from that do not involve higher
interest rates.
Fundamentally, the effect of budget deficits on interest rates is only the
proximate question that economists and policymakers wish to answer. The ultimate
question is where the resources come from to finance a deficit, and what
ramifications the use of those resources have for the nation’s welfare. A model in
which budget deficits do not affect interest rates still has consequences for the
nation’s welfare. In reviewing the major theories of the budget deficit’s effect, this
report explores the deficit’s effect on both interest rates and the nation’s welfare. It
begins with the mainstream view, and then explains how other versions differ from
that view.
Budget Deficits in the Mainstream View
In the mainstream economic view, budget deficits expand total spending
(aggregate demand), and thereby short-term economic growth.3 If a budget deficit
is the result of higher government spending, the additional government spending
expands aggregate spending directly. If a budget deficit is the result of tax cuts,
aggregate spending is expanded by an increase in spending by the tax cut’s
recipients.4 Note that the increase in the budget deficit described here is due to policy
changes, which is referred to as a change in the structural deficit. The change is not
due to changes in economic conditions, such as a fall in tax revenue due to a fall in
taxable income. When economic conditions cause tax revenues to fall, the actual
deficit would rise but the structural deficit would be unchanged.
In an economy at full employment, production (aggregate supply) cannot be
increased to match the increase in spending because all of the economy’s labor and
capital resources are already in use. This mismatch between aggregate demand and
aggregate supply must be resolved through market adjustment. Market adjustment
takes place in four distinct ways, each of which has consequences for interest rates.
! The greater demand for goods and services pushes up prices, leading
to a temporary increase in inflation. Because nominal interest rates
are equal to real (inflation-adjusted) interest rates and the expected
inflation rate, an increase in inflation (if anticipated) would push up
nominal interest rates.
3 For more information, see CRS Report RL31235, The Economics of the Budget Deficit, by
Brian Cashell.
4 For more information, see CRS Report RL30839, Tax Cuts, the Business Cycle, and
Economic Growth: A Macroeconomic Analysis, by Marc Labonte and Gail Makinen.
CRS-3
! If the Federal Reserve’s objective is to maintain a stable inflation
rate, as most observers believe, it will offset the rise in demand
through a contraction in monetary policy. It will increase short-term
real interest rates directly, and this will reduce interest-sensitive
spending (i.e., private investment and consumer durables).
! In the market for savings, the demand for savings has increased for
two reasons. First, because the government is now competing with
private firms for the same pool of national saving to finance its
deficit, the government has increased the demand for savings
directly, and interest rates rise as a result. Second, in response to the
greater demand for their goods, firms wish to increase their
investment spending in order to increase their supply of goods.
Investment demand increases in response to the desire for higher
production, and interest rates rise as a result.5
! In the money market, the increase in aggregate spending leads to an
increase in money demand. As a result, if monetary policy is left
unchanged, interest rates increase since the opportunity cost of
holding money is now greater. With higher interest rates, the same
supply of money can now be used to purchase more goods and
services because people do not hold on to money as long (money
velocity increases).
As a result of the interaction between these market forces, resources have been
shifted toward the government (or tax cut recipients) and away from the saving of the
private sector. This has consequences for the economy in the long run. In the long
run, economic growth is dependent on increases in private investment, productivity,
and the labor force. By identity, private investment equals national saving less the
budget deficit (foreign saving will be considered below). Thus, by reducing national
saving, budget deficits lead to less private investment. This reduces the size of the
economy in the long run, and future standards of living.6 If the deficit lasts for one
year, there would be a one-time reduction in growth. If the deficit lasted for several
years, it would reduce growth for the duration of that time.7
5 The conventional model usually assumes that the household savings rate would be
unaffected by a deficit-financed increase in government spending and that households would
save a fixed percentage of a deficit-financed tax cut. Since the average household saving
rate is low, it is usually assumed that the bulk of the tax cut would be spent. To the extent
that households save a portion of a deficit-financed tax cut, there will be less crowding out
(and stimulus) than a deficit-financed increase in government spending in the conventional
model.
6 If the deficit was used to finance public investment (e.g, highways), the effects on growth
would be different. In this case, private investment would still be reduced by an equivalent
amount but it would be shifted to public investment rather than public or private
consumption. Whether this increased or decreased economic growth would depend on
whether public investment was more or less productive than private investment.
7 Budget deficits could also cause interest rates to rise if they caused the debt to grow at
(continued...)
CRS-4
For this reason, economists often describe deficits as placing a burden on future
generations. Deficits allow individuals to enjoy more consumption today. But since
this higher consumption comes at the expense of lower saving, and hence lower
investment, it reduces the size of the economy in the future. Thus, individuals would
enjoy lower consumption in the future.8
The Conventional View With an Underemployed Economy
Thus far, it has been assumed that the expansion in the budget deficit was
taking place in a fully employed economy. Since all of the economy’s resources were
already in use, for the government to redirect resources to itself or tax cut recipients
required a reduction in the resources available to others. This reallocation occurred
through higher prices and interest rates, the latter of which “crowded out” private
investment and consumer durables.
But in an economy with underemployed labor and capital resources as a result
of a recession or low growth, the financing of a budget deficit is no longer a zero-sum
scenario in which any resources redirected to the government or tax cut recipients
must be fully offset by a reduction in the purchasing power of others. That is because
the increase in aggregate spending caused by the deficit leads to unemployed
resources being brought back into use, generating new aggregate production to match
the increase in aggregate spending. This is how expansionary fiscal policy
“stimulates” the economy during a recession. In an extremely underemployed
economy, enough unused resources are available to match the increase in aggregate
spending entirely. The increase in the budget deficit would be “multiplied” by the
fact that re-employed workers increase their spending as well, so that the total
increase in aggregate spending is larger than the increase in the budget deficit. In this
case, the budget deficit would be unlikely to have much of an effect on interest rates
and inflation, as long as it were eliminated once the economy returned to full
employment.9
The U.S. economy has not faced this extreme scenario since the Great
Depression. In all recessions since the 1930s, including the last one, some
underemployed resources have been available to be brought into use in response to
7 (...continued)
unsustainable rates, in which case investors would demand a risk premium to be induced to
hold government debt for fear of default. Budget deficits have never grown large enough
in the post-war period for this to be a relevant factor in the United States, where government
debt has always been considered riskless. Risk premiums have been observed for foreign
sovereign debt, however, particularly in the developing world. The crowding out argument
outlined in this section is not related to risk premiums.
8 For more information, see CRS Report RL30520, The National Debt: Who Bears Its
Burden?, by Marc Labonte and Gail Makinen.
9 In the previous section, one reason that budget deficits pushed up interest rates was
because the pool of national savings was fixed. In the extreme case of an underemployed
economy, budget deficits do not push up interest rates because the pool of national savings
grows along with national income. As deficits bring resources back into production, income
is generated, and some of that income will be saved, adding to national savings.
CRS-5
an increase in aggregate spending, but not enough so that there would not be any
increase in interest rates or prices within the framework of the conventional model.
Thus, expansionary fiscal policy would cause some increase in aggregate spending,
with larger increases the further the economy is from full employment. It would also
cause some increase in interest rates and prices, but less than in a fully employed
economy.10
If prices and markets do not adjust instantaneously and people’s expectations
do not change instantaneously, even at full employment it may be possible to
temporarily bring additional resources into production, thereby generating some
transient stimulus from an increase in the budget deficit. In other words, an increase
in the deficit is unlikely to ever lead to zero increase in aggregate spending in the
very short run. However, a deficit is likely to lead to a much smaller increase in
output and larger increase in prices at full employment than in a recession. And after
adjustment took place, the economy would return to full employment, with the same
higher interest rates and prices as described in the previous section. Economists tend
to disfavor attempts to push the economy above full employment in this way, since
the inevitable move back down to full employment could result in overshooting in
the opposite direction — recession.
Budget Deficits in The Mainstream View With
Capital Mobility
Thus far, the mainstream model described has used the assumption that the
United States is a closed economy, in which trade and capital does not flow between
it and the rest of the world. Relaxing this assumption has very important
ramifications for interest rates.
In a closed economy, the pool of savings available to the government to finance
its deficits and the private sector to finance its investment spending is fixed. For that
reason, any increase in the demand for that savings must push up interest rates. But
what if the firms and government of the United States could draw from the world
pool of savings, rather than being limited to the national pool of savings? If the
increase in the deficit were an insignificant fraction of world savings, it would have
no effect on world interest rates.11 Therefore, it would have no effect on U.S. interest
rates because U.S. interest rates, adjusted for risk, would have to equal world interest
10 For example, the unemployment rate fluctuated between 5.6% and 6% in 2002. If the full
employment rate of unemployment were 5%, as some economists have estimated, it would
suggest that the economy in 2002 was a little below full employment.
11 It is often argued that deficits should not affect interest rates because they are an
insignificant fraction of world wealth (or financial capital). This argument is not valid
because interest rates are determined by current saving and investment, not wealth, which
is a cumulative measure of past saving and investment, less depreciation. Although the
assumption that the increase in the deficit is an insignificant fraction of world saving is
made here for illustrative purposes, this assumption may be questionable for large deficits
in reality.
CRS-6
rates. Any time U.S. interest rates rose above (or fell below) world interest rates, in
theory capital would enter (or leave) the country seeking to take advantage of the
profit differential until the point where U.S. interest rates fell (rose) back to world
levels.12
Although with perfect capital mobility, a deficit would have no effect on interest
rates, a deficit would still have a cost to the economy. First, although the foreign
capital inflows permit a larger U.S. capital stock than would otherwise exist, the
returns from that capital would flow to foreigners rather than U.S. citizens.
However, U.S. workers would benefit from the larger capital stock in the form of
higher wages. The second cost comes from the effect the capital inflow has on the
economy. For foreign capital to enter the country, foreigners must buy U.S. dollars.
The increased demand for dollars causes the dollar exchange rate to appreciate. As
the dollar appreciates, U.S. exports become less competitive abroad and U.S. import-
competing firms become less competitive domestically. This causes the trade deficit
to expand, which reduces aggregate spending in the economy. In a world of perfectly
mobile capital, the expansion in the trade deficit would offset the expansion of fiscal
policy one-to-one, so that fiscal stimulus had no net effect on aggregate spending.
By adding capital mobility, crowding out has not been eliminated, it has been shifted
from the investment sector to the trade sector. Thus, although there is no effect on
interest rates, fiscal policy in this scenario no longer has any stimulative effect on
the economy.
The Response of Saving to Budget Deficits
and the Barro-Ricardo View
Saving in the Conventional View
In the conventional view, simple assumptions are made about the behavior of
private saving in response to a budget deficit. If the deficit is the result of a tax cut,
the conventional view assumes that the tax cut’s recipient would save a fraction of
that tax cut and spend the rest. Although there is no direct way to measure how much
of a tax cut recipients would be likely to save, since the average household saving
rate is very low (it averaged 4.4% of GDP in the 1990s and 1.0% of GDP in 2003),
it is typically assumed that tax cuts to individuals would be mostly spent. Therefore,
the rise in private saving would be much smaller than the fall in public saving and
little crowding out would be prevented.
12 Technically, this is referred to as an arbitrage condition. If interest rates did not equalize,
arbitrageurs could make unlimited profits by selling (buying) foreign securities and buying
(selling) U.S. securities. By attempting to exploit unlimited profit opportunities, the
arbitrageurs eliminate the interest rate differential.
This example also holds economic conditions in other countries constant. If other countries
interest rates rose or fell in step with ours (because they were pursuing the same policies,
for example), then there would be no interest rate differential and no resulting capital flow/
trade deficit.
CRS-7
“Supply siders” hold that the rise in private saving could be much greater than
would be implied by looking at the average saving rate. If saving is highly sensitive
to changes in interest rates, small increases in the after-tax rate of return (as a result
of lower marginal tax rates) could lead to large increases in saving (technically,
saving would be said to have a high and positive interest elasticity). If the interest
elasticity of saving were great enough, this would keep the budget deficit from
causing large changes in interest rates, and hence investment.13 Yet a look at the
long-run behavior of private saving in this country casts doubt upon the assertion that
small changes in tax rates could generate large changes in the private saving rate. The
corporate saving rate has been relatively constant in the post-war period, while the
household saving rate has dropped considerably in the past two decades. The
downward trend in the household saving rate has occurred during a period with many
changes in marginal tax rates, including large reductions in top marginal income tax
rates, marginal tax rates on investment income, and the expansion of tax-preferred
savings vehicles in the 1980s and the 2000s.14
It is also worth considering, from a theoretical perspective, that private saving
could be sensitive to interest rates in the opposite way. Rather than saving more in
response to higher interest rates, individuals could save less. If individuals are
primarily target savers, saving to meet a goal such as owning a house, car, or to
support a certain standard of living in retirement, higher interest rates would make
that goal easier to meet. This would cause them to save less in response to a budget
deficit, making the deficit’s negative effect on long run growth even greater than
under the conventional view. Thus, even if incentives are an important determinant
of saving behavior, it is not clear in which direction the incentives cause behavior to
react. And there is no straightforward evidence to suggest which view is correct.
Indeed, the rapidly rising stock market in the late 1990s coincided with a decline in
the household savings rate to nearly zero, as the target-saver view would predict.
Saving in the Barro-Ricardo View
Are there explanations that suggest a larger private saving response to a budget
deficit? One weakness in the mainstream view is that the explanation of how savers
react to changes in the government’s fiscal position is not well developed. There is
no explanation of how today’s policy decisions affect the future, and how individuals
incorporate their perceptions of the future into their plans today. The Barro-Ricardo
view, named after the 19th century economist David Ricardo, and Robert Barro who
revived and developed Ricardo’s theory, addresses this issue.15 Based on a very
13 It should be noted that if this view is correct, then tax cuts would be significantly less
stimulative than government spending when the economy is below full employment. That
is because the portion of a tax cut that is saved is not contributing to an increase in aggregate
spending. For more information, see CRS Report RS21136, Government Spending or Tax
Reduction: Which Might Add More Stimulus to the Economy?, by Marc Labonte.
14 For more information, see CRS Report RL30873, Saving in the United States: How Has
It Changed and Why Is It Important?, by Brian Cashell and Gail Makinen.
15 This view is also referred to as “Ricardian Equivalence.” Robert Barro, “Are Government
Bonds Net Wealth?,” Journal of Political Economy, vol. 82, no. 6 (Nov/Dec 1974), p. 1095.
(continued...)
CRS-8
particular set of assumptions, Barro showed how deficits could have no effect on
interest rates. Barro assumed that people were perfectly rational, planned their
lifetime consumption through optimization, had infinite life spans (he suggested that
concern for offspring on par with one’s own well-being could substitute for infinite
life spans in reality), could borrow against future earnings without limit, and were all
taxed equally. If the deficit finances government purchases, those purchases must be
perfect substitutes for private consumption. He assumed that if the deficit finances
tax cuts, the tax cut was lump-sum in nature, so that incentives to work or save are
not changed.
Under these circumstances, he reasoned, individuals would know that any
increase in the budget deficit would have to be offset by an increase in their tax
burden or decrease in their government services in the future.16 In light of the larger
deficit, their previous lifetime consumption plan would lead to too much
consumption today and too little in the future. To make up for the reduction in future
consumption brought about by the deficit, they would save more now. Thus, as the
government placed greater demand on the nation’s savings, the pool of savings would
expand, so that no upward pressure was placed on interest rates. Since interest rates
and investment levels are the same, long-term growth would be the same. This also
means that an increase in the budget deficit would have no short-run stimulative
effect on the economy since national saving (and thus aggregate spending) has stayed
the same — the stimulus provided by the government has been offset by a contraction
in private spending.
As can be seen, the fact that interest rates do not rise in the Barro-Ricardo view
does not mean that deficits can be financed without using any resources. Instead of
transferring resources from the investment sector (as in the conventional view), or the
trade sector (as in the capital mobility view), the resources are transferred from
private consumption. In this sense, the effect of a deficit under the Barro-Ricardo
view is similar to a tax increase: resources are transferred to the government through
a reduction in private consumption.
The Barro-Ricardo view has a clear advantage over the conventional view: it
offers an explanation for how individuals adjust their saving behavior to take into
account fiscal policy. But the strict conditions adopted to make this explanation
tractable are also its greatest weakness. If any of the highly exacting prerequisites are
weakened, the results may no longer hold. For example, what if individuals do not
plan out their lifetime consumption because of uncertainty or myopia? What if
15 (...continued)
For an overview of the issues raised in the debate on the Barro-Ricardo Theory, see John
Seater, “Ricardian Equivalence,” Journal of Economic Literature, vol. 31, no. 1 (Mar.
1993), p. 142.
16 In a later paper, Barro drew a distinction between anticipated and unanticipated increases
in the budget deficit. He reasoned that the Barro-Ricardo view of zero effect of deficits on
the economy and interest rates held for anticipated increases in the budget deficit.
Unanticipated increases, by contrast, could have real effects on the economy and interest
rates. Robert Barro, “Federal Deficit Policy and the Effects of Public Debt Shocks,”
Journal of Money, Credit, and Banking, vol. 12, no. 4 (1980), p. 747.
CRS-9
individuals do not understand the link between deficits and their future welfare? Or
what if individuals, who in real life are not all the same, think that someone else will
get stuck with the bill? For example, nobody’s lifespan is infinite and not everyone
has children. Under any of these cases, individuals would no longer increase their
saving enough to match the increased demands placed on national saving by the
budget deficits. Therefore, deficits would cause interest rates to rise.
Empirical Evidence
Comparing changes in budget deficits to changes in interest rates is not a valid
way to determine whether budget deficits affect interest rates. That is because there
are many other factors that affect interest rates. To determine the effect of budget
deficits on interest rates, one must hold these other factors constant using statistical
methods. Otherwise, the effect of budget deficits on interest rates could be mis-
estimated or even reversed. For example, interest rates often fall during recessions
since investment demand is reduced and precautionary saving may rise. If the
business cycle has a greater effect on interest rates than budget deficits, and changes
in the budget deficit (more specifically, changes in the structural deficit) happen to
coincide with recessions, a simple comparison of deficits and interest rates would
suggest that deficits cause interest rates to fall. In reality it is the recession that is
causing interest rates to fall, and after controlling for the recession, very different
results may be achieved.17
It should also be stressed that empirical evidence that budget deficits do not
affect interest rates does not demonstrate that government budget deficits are
burdenless. Empirical tests of the effects of deficits on interest rates are only tests
of the conventional view without capital mobility. The capital mobility view and
Barro-Ricardo view both explain how deficits could have no effect on interest rates,
and yet the deficits still impose a burden in both views. In the capital mobility view,
they crowd out the trade sector of the economy; in the Barro-Ricardo view, they
crowd out current private consumption.
Over the past two decades, there have been dozens of econometric studies
estimating the effects of budget deficits on interest rates. The studies have reached
very different conclusions because they have covered different periods, included
different variables, and modeled the relationship between the variables and interest
rates differently. Seemingly insignificant changes in the structure of the model, such
as how long a deficit should affect interest rates, how deficits and interest rates are
defined, and the mathematical form used to represent the relationship can drastically
change the results. Not only do variations in the model used lead to different results,
but some researchers make assumptions at odds with theory. These include the use
of the actual budget deficit rather than the structural deficit, the effect of debt on
interest rates instead of the deficit, the use of nominal instead of real interest rates,
the omission of foreign capital flows, and so on. Another problem with the statistical
17 This would be similar to claiming that men are fatter than women because on average men
weigh more than women, without controlling for the fact that men are on average taller than
women.
CRS-10
method used in all of the studies reviewed, time series analysis, is that there can be
no fundamental change in the economic regime over the time period for the results
to be valid. Yet many studies commingled data from both the fixed and floating
exchange rate period; the period in which capital mobility has been relatively free
and the period that used capital controls; the 1970s, in which the Fed arguably
tolerated high inflation, and the period since, when it did not. Some studies even
included World War II, although this period featured the rationing of goods and
credit, price controls, and the fixing of interest rates by the Federal Reserve.
Rather than review all of these studies, this report will summarize the results of
three literature reviews. In 1991, Barth et al. published a literature review of 42
papers exploring the link between deficits and interest rates.18 They found the
evidence to be mixed, with 17 studies finding deficits to have a positive and
statistically significant effect on interest rates, 6 having mixed effects, and 19 having
statistically insignificant or negative effects. Some studies with a positive
relationship estimated that the effect of a deficit equal to 1% of GNP could increase
interest rates as much 0.4-1.2 percentage points. Summarizing the studies, the
authors reach a number of conclusions:
! More of an effect is found on long-term interest rates than short-term
rates. This makes sense since overnight rates are targeted by the
Federal Reserve, and thus heavily influenced by monetary policy in
the short run.
! Changes in structural budget deficits have more of an effect on
interest rates than changes in the actual budget deficit. The
structural (full-employment) deficit is a measure of the deficit that
eliminates the effects of the business cycle on the budget. A
downturn in economic activity automatically causes revenues to fall
as taxable income falls and certain types of government
expenditures, such as unemployment benefits, to rise. Using the
actual budget deficit in a statistical study would skew the results
because the actual deficit is being determined in part by the same
factors (the state of the economy) that are determining interest rates.
Using the structural budget deficit or advanced statistical techniques
avoids this problem.
! A link between deficits and interest rates is less likely to be found
when high frequency (monthly or quarterly) data are used. This
makes sense given how “noisy” (volatile) high frequency data tend
to be.
! An effect is more likely to be found if a study uses a measure of the
expectation of future deficits rather than contemporaneous deficits.
This makes sense if investors are rational and forward looking —
18 James Barth et al., “The Effects of Federal Budget Deficits on Interest Rates and the
Composition of Domestic Output,” in Rudolph Penner, ed., The Great Fiscal Experiment
(Washington DC: Urban Institute, 1991), p. 69.
CRS-11
investors are not only concerned about what will happen today, but
also what will happen in the future. However, the proper measure
of future deficits is controversial. The CBO or OMB baseline is not
a proper measure since it is not a “best guess” of future policy.
! The link between deficits and interest rates is better established
when the deficits finance government spending than when they
finance tax cuts. The effect of a tax cut on interest rates would be
expected to be smaller if tax cut recipients save some of the proceeds
of the tax cut.
! Studies using a statistical method called vector-autoregression
consistently found deficits to have no effect on interest rates.
Basically, the vector-autoregression uses a small number of variables
and makes no attempt to identify the theoretical relationship between
the variables. This method has advantages and disadvantages — it
would not be as accurate as an estimation of the correct theoretical
model, but given the lack of consensus over what the correct model
is, it may be more accurate than an incorrect theoretical model. It is
not a method that can control for investors’ expectations of future
deficits, which is a drawback if this an important factor.
They conclude that
Since the available evidence on the effects of deficits is mixed, one cannot say
with complete confidence that budget deficits raise interest rates and reduce
saving and capital formation. But equally important, one cannot say that they do
not have effects.19
This conclusion seems apt given the balanced range of results in the studies they
surveyed. They also look at studies that attempt to estimate whether deficits crowd
out investment, the trade sector, or consumption, and find the results of these studies
to be mixed as well.
A Joint Economic Committee study surveyed many of the same studies.20 It
surveyed five articles, all of which were included in the Barth survey, on the
relationship between deficits and interest rates, all five of which found the
relationship to be negative or statistically insignificant. It also surveyed 15 articles
on the relationship between deficits and consumption, which the study took to be a
test of the Barro-Ricardo theory, and found that seven studies supported the Barro-
Ricardo theory and eight rejected it.
19 Ibid, p. 94.
20 Robert O’Quinn, Fiscal Policy Choices: Examining the Empirical Evidence, Joint
Economic Committee, (Washington, DC: Nov. 2001).
CRS-12
Gale and Orszag review journal articles written on the subject in the 1990s,
since the Barth paper was published.21 They find that of the 16 studies published
since the Barth paper, 11 found that deficits had a positive and statistically significant
effect on interest rates, five had mixed effects, and none found statistically
insignificant or negative effects. This evidence is quite a bit stronger than the
evidence found in the Barth review, and they attribute the change in outcome to some
changes in research techniques, which they believe to be superior to earlier studies.
(The studies also covered several more years of data.) Specifically, they believe the
more recent studies to be superior because all of the recent studies focus on changes
in long-term budget forecasts rather than current-year deficit totals and because they
differentiate between long-term interest rates and short-term interest rates. These
developments seem consistent with the expectations theory of interest rates, which
suggests that long-term interest rates are determined by a combination of today’s
short-term interest rates and investors’ perception of future short-term rates. Long-
term deficits would therefore have a greater effect on interest rates than a temporary
deficit. Short-term rates may also offer misleading evidence since they are so heavily
influenced by the Fed, which targets overnight interest rates.
Gale and Orszag also examined the effects of a budget deficit on interest rates
in 10 leading macroeconomic forecasting models, including the models of DRI,
Macroeconomic Advisers, CBO, and the Federal Reserve. In all 10 models, a budget
deficit equal to 1% of GDP would increase interest rates, with a range of 0.1 to 1.0
(mean=0.52) percentage points after one year and 0.05 to 2.0 (mean=0.99) percentage
points after 10 years. It is striking that all forecasters, with their reputations and
profits at stake in producing the most accurate results, use models in which deficits
increase interest rates.
Conclusions
Any explanation of the budget deficit-interest rate relationship must first come
to grips with an indisputable fact: budget deficits use real resources. When the
government borrows from the public to finance public spending or tax cuts, the
resources must come from somewhere. In mainstream economic theory, the
resources come from the nation’s pool of saving, which pushes up interest rates for
simple supply and demand reasons. This “crowds out” private investment which was
competing with government borrowing for the same pool of national saving. For this
reason, economists often describe deficits as placing a burden on future generations.
But other theories offer different explanations of where the resources come from
that do not involve higher interest rates. In the capital mobility view, foreigners lend
the United States the savings it needs to finance a deficit, leaving interest rates
unaffected. But as foreign capital comes to the country, the dollar must appreciate.
This causes U.S. exports and import-competing industries to become less competitive
and the trade deficit to expand. In the Barro-Ricardo view, forward-looking, rational,
infinitely-lived individuals see that a budget deficit would result in higher taxes or
21 William Gale and Peter Orszag, “The Economic Effects of Long-Term Fiscal Discipline,”
Urban Institute-Brookings Tax Policy Center Discussion Paper, Dec. 2002.
CRS-13
lower government spending in the future. Therefore, they reduce their consumption
and save more today. This provides the government with the saving needed to
finance its deficit, placing no upward pressure on interest rates. Although
theoretically compelling, the very particular assumptions made in the Barro-Ricardo
view lead one to question its practicality as an explanation of the real world.
Fundamentally, the effect of budget deficits on interest rates is only the
proximate question that economists and policymakers wish to answer. The ultimate
question is where the resources come from to finance a deficit, and what
ramifications the use of those resources have for the nation’s welfare. Empirical
evidence that budget deficits do not affect interest rates is not evidence that
government budget deficits do not impose a burden. The capital mobility view and
Barro-Ricardo view both explain how deficits could have no effect on interest rates,
and yet still impose a burden in both views. In the capital mobility view, they crowd
out the trade sector of the economy; in the Barro-Ricardo view, they crowd out
current private consumption. Ironically, the conventional view is the only one which
suggests that deficits can stimulate aggregate spending in the short run, so
policymakers attempting to stimulate the economy should hope to see interest rates
rise.
Simply comparing changes in budget deficits to changes in interest rates is not
a valid way to determine whether budget deficits affect interest rates because there
are many other factors that simultaneously affect interest rates. These include the
state of the economy, the saving behavior of individuals and companies, the
investment opportunities of business, the state of financial markets, demographics,
and the financial relationship between the United States and the rest of the world. To
determine the effect of budget deficits on interest rates, one must hold these other
factors constant using statistical methods. Otherwise, the effect of budget deficits on
interest rates could be misestimated or even reversed.
But controlling for other factors requires a model that explains how a deficit and
those other factors affect interest rates. Thus, even those who are skeptical of
theoretical explanations are dependent on theoretical models to glean the relationship
from the empirical evidence. Because so many different models of the economy
exist, the empirical evidence is mixed, with some models offering positive evidence
about the deficit-interest rate relationship while others offer evidence refuting the
relationship. Some studies are questionable because they make assumptions at odds
with the underlying theory (e.g., measuring the relationship of interest rates and debt,
rather than deficits). Gale and Orszag (2002) argue that more recent evidence tends
to find a stronger, positive relationship between the two, with all sixteen of the
studies they surveyed finding positive or mixed evidence. In addition, 10 major
forecasting models all predict that a budget deficit would increase interest rates.
According to Gale and Orszag (2002), the models predict that a budget deficit equal
to 1% of GDP would increase interest rates, with a range of 0.1 to 1.0 (mean=0.52)
percentage points after one year and 0.05 to 2.0 (mean=0.99) percentage points after
10 years.