Order Code RS21136
Updated March 13, 2002
CRS Report for Congress
Received through the CRS Web
Government Spending or Tax Reduction:
Which Might Add More Stimulus to the
Economy?
Marc Labonte
Economist
Government and Finance Division
Summary
Members have proposed several “stimulus” measures to help end the recession. A
fundamental difference in competing stimulus packages is how much of the stimulus
should be devoted to government spending and how much should be devoted to tax cuts.
This report considers that issue in the context of conventional economic analysis. It first
identifies any policy change that increases the budget deficit (or reduces a surplus) and
is not entirely saved by the recipient as “stimulative” if the economy is operating below
its full potential. It then separates the short-run effects of a budget deficit from the long-
run effects. In this context, certain spending proposals may be more stimulative than
certain tax reductions in the short run if they result in a bigger boost in aggregate
spending. This advantage may come at the cost of forgone growth in the long run,
however. The stimulative effects of two specific types of spending proposals are
analyzed, an extension and expansion of unemployment benefits and health care subsidies
for unemployed workers. P.L.107-147 was signed into law on March 9, 2002. This
report will not be updated.
In March 2001, the U.S. economy entered a recession, and Members of Congress
have since proposed several “stimulus packages” in response. On March 9, 2002 a
stimulus package was signed into law (P.L.107-147) that included an acceleration in
depreciation rates and an extension of unemployment benefits. Throughout the debate,
there was heated debate concerning how to get the most short-term “bang for the buck”
out of any stimulus package, and one key element of that debate is whether spending
proposals or tax reductions would be more effective in that regard. To answer that
question, one must first understand why the economy has entered a recession. While the
ultimate causes are complex and diverse, they can be boiled down to one proximate cause:
spending (aggregate demand) in the economy is currently lower than what could be
produced if the economy were at full employment (aggregate supply).
Given the cause of the recession, how should a stimulus proposal be judged? By
whether it boosts economic growth might seem to be an obvious answer, but the answer
is a little more complicated. Policy proposals have distinct short-run and long-run
Congressional Research Service ˜ The Library of Congress

---

CRS-2
economic effects, and often a proposal beneficial in the short run may be detrimental in the
long run, and vice versa. In the short run, many proposals would affect the economy by
boosting aggregate spending; in a recession, their effects on the productive capacity
(“supply side”) of the economy are likely to be slight because spending is not adequate to
support existing production. In the long run, proposals can affect economic growth by
changing individuals’ work effort, productivity, and the saving rate. By contrast, a
proposal’s effect on aggregate spending can be ignored in the long run because aggregate
spending can never be permanently altered; once prices, wages, and credit markets adjust,
aggregate spending effects are dissipated. If one is to judge a proposal whose aim is to
stimulate the economy during a recession, it seems logical to judge it on the basis of how
much it boosts aggregate spending in the short run. This does not alter the fact that, for
better or worse, it will also have long-run effects.
Evaluating the stimulative nature of different proposals does not answer the question
of whether the fiscal stimulus for 2002 already in place, which includes a $44 billion
increase in discretionary spending over the baseline and $31 billion in newly phased-in
provisions of last year’s major tax act (P.L.107-16), is sufficient to revive the economy.
Nor does this evaluation constitute an endorsement of the need for a fiscal stimulus.
Expansionary monetary policy (i.e., lower interest rates) by the Federal Reserve also
stimulates the economy. In addition, prices, wages, and credit markets naturally adjust to
bring aggregate spending back into line with aggregate production. A look at the
historical record suggests that these two factors have proven highly capable of moving the
economy out of recession in the absence of a fiscal stimulus.1
Short-Run Effects
As a first approximation, any proposal – whether it be a tax or spending proposal –
that increases the size of the budget deficit (or decreases the size of a surplus) would be
expected to boost aggregate spending by some degree, thereby stimulating an economy
operating below its full potential.2 The key is that current proposals would be financed
with borrowed money, and credit markets would not fully adjust immediately in terms of
higher interest rates, so that the spending is “new.” In that context, if the proposal is offset
by spending cuts in other parts of the budget or other tax increases, it will largely lose its
short-run effect on aggregate demand. Proposals labeled as a “stimulus package” or, say,
“higher military spending” would both stimulate aggregate spending; the most important
determinant of a proposal’s effect is how much the budget deficit increases as a percentage
of gross domestic product (GDP). Because the policy brings unused resources back into
production, the increase in aggregate spending will be greater than the budgetary cost of
the proposal by some “multiplier.” To the extent that interest rates rise as a result of the
decline in saving (increased budget deficit), the multiplier will be reduced because the
proposal will tend to “crowd out” private investment and other interest-sensitive spending.
Economic theory suggests it will also crowd out exports and import-competing goods
because higher interest rates will tend to attract foreign capital that will cause the exchange
value of the dollar to appreciate. Similarly, if the proposal causes inflation to rise, some
1 See CRS Report RL31237, The Current Economic Recession: How Long, How Deep, and How
Different from the Past?, by Marc Labonte and Gail Makinen.
2 For a more detailed analysis, see CRS Report RL31235, The Economics of the Federal Budget
Deficit, by Brian Cashell.

---

CRS-3
of the spending will be crowded out as real income falls. More crowding out will occur
as production moves closer to its full potential.
All measures that increase the size of the deficit will boost aggregate spending in a
recession to some extent, but some measures are likely to boost it more than others. The
government can increase the budget deficit (or reduce the surplus) in three ways: it can
spend money directly by purchasing goods and services, it can make income transfers by
giving cash or in-kind benefits to selected individuals, or it can cut taxes. (Note that while
in budgetary terms government spending and government transfers are both defined as
outlays, they have different economic meanings and effects.) Government spending is
likely to boost aggregate spending more than tax cuts or a government transfer because
some part of a tax cut or transfer would be saved. By definition, any resources that are
saved cannot add to aggregate spending. Similarly, different tax cuts and transfers will
have different short-run effects because they offer different incentives to save or spend and
their recipients have different spending/saving patterns.
Who should a tax cut or government transfer target to get the most short-run “bang
for the buck?” Empirically, there is evidence that lower income recipients are likely to
spend more of a tax cut or transfer payment than higher income recipients. There is no
strong theoretical rationale to explain this. One fairly well-accepted theoretical
explanation holds that lower income recipients are more likely to be “liquidity-constrained”
than higher income recipients, meaning that they are more likely to lack access to credit
markets that allows them to alter their borrowing or saving at will. Thus, spending a tax
cut or transfer payment may allow them to achieve a level of consumption that they would
have undertaken if they had unrestrained access to credit markets.
If the economic effects of both transfer payments and tax cuts depend on the
recipient’s behavioral response, is there any reason to generally expect one or the other to
cause a greater boost in aggregate demand? To answer this question, two further points
are worth considering. First, about 20% of low-income households accrue no federal
income tax liability, although some do pay federal payroll and excise taxes. Thus, there
are more options for targeting lower income recipients with transfer payments. Second,
it would be difficult to target an individual tax cut in such a way that it could only be
spent.3 The recipients of tax rebates, income tax cuts, or payroll tax holidays would be
free to decide whether to save or spend their payment. A sales tax holiday is a notable
exception, but raises problematic administrative and federalism issues since sales taxes are
levied by state governments. By contrast, there are other types of tax cuts such as the
expansion of tax deferred savings accounts that are intended to stimulate more saving, and
hence less spending. Certain types of transfer payments, such as food stamps or rental
vouchers, are more easily targeted in such a way that they cannot be saved, and must be
spent. Even these policies do not prevent individuals from shifting an offsetting amount
of any other income into saving, however. Thus, government spending on goods and
services remains the only way to prevent the short-term increase in aggregate spending
from being partially offset by private saving.
3 It is easier to target a business tax cut in such a way that it must be spent. For instance, a
business could claim the proposed accelerated depreciation only if it purchased new capital. See
CRS Report RL31134, Using Business Tax Cuts to Stimulate the Economy, by Jane Gravelle.

---

CRS-4
In determining a proposal’s short-run efficacy, another issue to consider is the lag in
policy implementation (a separate issue from the lag caused by the legislative process).
It may take a considerable amount of time to set up an administrative infrastructure before
certain government spending or transfer proposals can be implemented. Alternatively,
other proposals expand existing programs and, in some cases, can be implemented quite
quickly. Tax cuts typically take effect in the following fiscal year, but exceptions can be
made, as was the case with the 2001 tax “rebate.” The rebate still entailed a lag: P.L.107-
16 was signed into law June 7, 2001, the first rebate checks were issued the week of July
23, and the last rebate checks were issued the week of September 24.4
Long-Run Effects
Proposals that are good for the economy in the short run, in terms of boosting
aggregate spending, are often bad for the economy in the long run due to the spending-
saving dichotomy. Long-run growth depends on investment, work effort, and productivity
gains. Investment, in turn, depends on saving. Since a budget deficit lowers national
saving (while a surplus increases national saving), virtually any policy that increases the
budget deficit would lower national saving. Thus, measures that boost aggregate
spending in the short run will lower saving, and therefore growth, in the long run. The
extent that national saving is reduced, in turn, depends on how much private saving offsets
the reduction in public saving (the increase in the deficit).
The fact that some part of a tax cut is likely to be saved gives tax cuts an advantage
over most government spending in the long run. In turn, a tax cut can be crafted to
generate more saving by targeting higher income individuals (who have higher saving
rates) or targeting saving directly (e.g., expanding tax-deferred saving accounts). Whether
tax cuts actually induce private saving depends on whether individuals react to the higher
after-tax rate of return by saving more because it is more rewarding (the substitution
effect) or by saving less because less saving is now needed to reach their goal (the income
effect). Similarly, tax cuts can be crafted to increase the rewards for working. Whether
or not this generates more work effort depends on whether people work more in response
to higher after-tax pay (substitution effect), or work less because less work is now needed
to achieve a given standard of living (income effect). If a tax cut generates more work
effort, this could also help offset the negative effects on economic growth of raising the
deficit.5
By contrast, most government spending and transfers do not increase the incentives
for individuals to work or save. The negative effects on long-run growth caused by the
larger budget deficit, therefore, would not be offset. There are exceptions such as
government spending on public investment. Instead of diverting resources from private
investment to public consumption, as most government spending does, it diverts resources
from private investment to public investment. Whether this adds or detracts from long-run
growth depends on whether public or private investment, respectively, has a higher rate
4 For more information, see CRS Report RS20926, Tax Rebate for 2001: Frequently Asked
Questions, by Louis Talley; and CRS Report RS20939, The Rate Reduction Tax Credit (“Tax
Rebate”) in P.L.107-16, by Gregg Esenwein and Steven Maguire.
5 For more information, see CRS Report RL30839, Tax Cuts, the Business Cycle, and Economic
Growth, by Marc Labonte and Gail Makinen.

---

CRS-5
of return. Productivity in government may grow more slowly than it does in the private
sector over time, also detracting from long-run growth.
Although transfer payments may have significant social value, they can also reduce
the incentives to work and save if they are means-tested for income level. A reduction in
transfer payments when income rises is equivalent to a marginal income tax, and some
individuals may reduce their work effort or saving to avoid this de facto marginal tax.
Likewise, unemployment benefits can lower the incentive to find work. On the other hand,
there are examples of transfer payments that increase the incentive to work, such as
transfer payments for which only employed individuals are eligible.
Most of the spending and transfer proposals associated with the stimulus bill are too
small as a percentage of GDP to have measurable macroeconomic effects. There are,
however, two exceptions, extending or expanding unemployment benefits and health care
subsidies to the unemployed, and these will be evaluated in the following sections.6
Economic Effects of Unemployment Insurance Proposals
There have been many recent proposals to temporarily extend or expand
unemployment insurance benefits. In terms of short-run economic effects, an extension
of eligibility for unemployment insurance benefits seems likely to be spent by recipients by
both of the criteria listed above, thus offering a high short-term “bang for the buck” to
aggregate spending. First, the average recipient of unemployment benefits has a below-
average income, and would thus be more likely to spend the benefits. Second, recently
unemployed individuals are likely to be particularly susceptible to “liquidity constraints,”
as they may be unable to replace as much of their lost income as they would desire by
borrowing or drawing down their savings. For liquidity-constrained individuals, spending
unemployment benefits would offer a way to prevent their consumption levels from falling
to an undesirable extent in response to their loss of income.
In recessions, most job losses fall under the category of “involuntary” unemployment
– there are not enough jobs available given the number of job seekers. Thus, in the short
run an extension or expansion of unemployment benefits during a recession is not likely
to detract significantly from the incentives to work. Since many of the proposals are
temporary – assuming they are not extended – there should be no long-run impact on the
desire to work and no long-run negative budgetary impact, and hence no significant effect
on long-run economic growth. Those proposals that are permanent could have some
negative impact on the incentive to work when there is not a dearth of available jobs in the
economy, and could thus reduce long-run growth through the labor supply channel.
Permanent policy changes would also have a permanent negative impact on the budget
balance, permanently reducing national saving and, hence, long-run growth.
There would be some lag in policy implementation from an administrative
perspective, particularly for proposals that update work history information or expand the
pool of eligible workers (e.g., expand coverage to part-time workers). In addition to the
6 For a detailed description of proposals, see CRS Report 95-742, Unemployment Benefits:
Legislative Issues in the 107th Congress, by Celinda Franco; and Health Insurance for Displaced
Workers, by Julie Stone and Jean Hearne, in the CRS Terrorism Electronic Briefing Book..

---

CRS-6
administrative lag, proposals to extend the period of benefit eligibility entail another lag.
A proposal to extend unemployment benefits from, say 26 weeks to 39 weeks, would only
increase aggregate spending beginning 27 weeks after any given worker was unemployed.
Economic Effects of Expanding Health Coverage for the
Unemployed
There have been proposals to subsidize health care premiums for qualifying
unemployed workers, either through direct payments (for example, the Democratic
substitute to H.R.3090) or an individual income tax credit. Since the tax credit in
H.R.3529 is refundable, advanceable, and payable directly from the government to health
insurance companies, the analysis is the same as in the case of a direct payment.
These proposals would add to aggregate spending in the short run if individuals
receiving the benefit did not increase their saving by an offsetting amount. Are the
recipients likely to offset the benefit by increasing their saving? It can be argued that a net
increase in spending would occur due to liquidity constraints. But it is harder to argue that
a net increase in spending would occur because the benefit would mostly be claimed by
low-income recipients. Even with a large subsidy, health care premiums would likely cost
hundreds of dollars a year on average, arguably placing many proposals out of the range
of lower income unemployed individuals. Major provisions of both bills are targeted to
workers who were insured before becoming unemployed, and these workers also tend to
have higher incomes.7 The stimulus effects would likely be slowed by an implementation
lag while the government determined eligibility and approved payments.
Many of the proposals are temporary and, therefore, would have negligible economic
effects in the long run. If the proposals were made permanent, they would have a negative
impact on the budget deficit and national saving, lowering long-term growth. If
maintaining health insurance is unaffordable for the long-term unemployed, the proposal
should not reduce the incentive to work. In that case, the proposal’s only long-run
economic effect could be to increase the incentive to change jobs, which would have an
ambiguous effect on growth. One effect of greater job turnover would be to generate
more periods of “frictional” or “between job” unemployment overall, which would result
in less overall labor supply and lower long-run growth. But another effect of greater
turnover might be better “job matching” between employees and employers, which in turn
would lead to higher productivity, and, hence, higher long-run growth.
7 One of the major channels through which both bills would operate is the COBRA program. A
1999 survey found that only 32% of low-income households (incomes below 200% of the poverty
level) would be eligible to maintain their health care through the COBRA program, compared to
67% of high-income households (incomes above 300% of the poverty level). It also found that
among the unemployed, only 5% of low-income households were covered by COBRA, compared
to 11% of high-income households. In the Senate version of H.R.3090, states could subsidize the
portion of COBRA benefits not covered by the federal government and offer health insurance to
unemployed workers ineligible for COBRA through Medicaid. However, states would have to
partly finance these proposals themselves at a time when many states are facing budget shortfalls.
See Stephen Zuckerman, Jennifer Haley, and Matthew Fragale, Could Subsidizing COBRA Heath
Insurance Coverage Help Most Low-Income Unemployed?, Urban Institute, October 17, 2001.
For a description of the COBRA program, see CRS Report RL30626, Health Insurance
Continuation Under the COBRA Program, by Heidi Yacker.

---