Microsoft Word - wp100Blinder.doc

The Case Against the Case Against

Discretionary Fiscal Policy

Alan S. Blinder

Princeton University

CEPS Working Paper No. 100

June 2004

Prepared for Federal Reserve Bank of Boston conference, “The Macroeconomics of Fiscal

Policy,” Chatham, Mass., June 14-16, 2004. I am grateful to Luke Willard for research assistance,

to Kathleen Hurley for handling the manuscript, and to Princeton University’s Center for

Economic Policy Studies for financial support.

Serious discussion of fiscal policy has almost disappeared.

— Robert M. Solow (2002)

Times change. When I was introduced to macroeconomics as a Princeton University

freshman in 1963, fiscal policy—and by that I mean discretionary fiscal stabilization

policy—was all the rage. The policy idea that would eventually become the Kennedy-

Johnson tax cuts was the new, new thing. In those days, discussions of monetary policy

often fell into the “oh, by the way” category, with a number of serious economists and

others apparently believing that monetary policy was not a particularly useful tool for

stabilization policy.

The appropriate role for central bank policy was often said to be

“accommodating” fiscal policy, which was cast in the lead role.

Thus many people,

probably including President Kennedy, thought that Walter Heller, who was then

chairman of the Council of Economic Advisers, was more instrumental to stabilization

policy than William McChesney Martin, who was then chairman of the Federal Reserve

Board. Indeed, it was said that Kennedy only remembered that Martin was in charge of

monetary policy by the fact that both words began with the letter M.

Multiply by –1, and you have a capsule summary of the conventional wisdom today.

As the opening quotation suggests, virtually every contemporary discussion of

stabilization policy by economists—whether it is abstract or concrete, theoretical or

practical—is about monetary policy, not fiscal policy. It never crosses anyone’s mind that

Greg Mankiw might be more influential in formulating stabilization policy than Alan

See, among others, Stein (1969, pp. 36-352).

Curiously, this phrase has survived into current Federal Reserve jargon. Even though the Fed is clearly in

the driver’s seat when it comes to stabilization policy, it routinely refers to increasing or reducing the

degree of “monetary accommodation.”

Stein (1969, p. 4). I was astonished to find, in researching this paper, that the index of Heller (1966) does

not contain a single reference to Martin.

Greenspan. And President Bush, I trust, does not need a mnemonic to remember what

Greenspan does for a living. This paper explores whether this complete about-face in the

conventional wisdom was well justified.

But don’t be alarmed by the title.

It speaks only of “the case against the case against”

fiscal policy, not “the case for.” I have no intention, at a Federal Reserve conference or

anywhere else, of challenging the now-standard view that the central bank should and

does have a dominant role in stabilization policy. This paper agrees that a sharp revision

of the naively optimistic views held by some economists circa 1966 was called for. But

it suggests that the pendulum may have swung just a bit too far, that the case against

fiscal policy may have been taken to extremes. Yes, monetary policy merits the

preeminent role in stabilization policy that it now holds. It is perfectly appropriate for

there to be 10-20 conferences on monetary policy for every one on fiscal policy. My

modest suggestion is only that the idea of using fiscal policy to help stabilize demand

should not be relegated to the dustbin of history. There are circumstances under which

the lessons of Lord Keynes are best not forgotten.

I. The Issues

The prevailing view today is that stabilization policy it is about filling in troughs and

shaving off peaks, that is, reducing the variance of output around a mean trend that is

itself unaffected by monetary or fiscal policy. For the most part, I will adhere to this

canon. But note that contemporary conventional wisdom makes two assumptions that are

at least debatable:

Old-timers may note that I have adapted the title from Solow (1966), and for much the same reasons that

he used it.

Assumption 1: The macroeconomy is not subject to hysteresis. In a system with a unit

root, any shock to aggregate demand—whether it be from fiscal policy or anything else—

will leave a permanent impact on output. There are many possible rationales for

hysteresis in a macroeconomic setting. One example is insider-outsider models (Lindbeck

and Snower (1988)) in which workers who become unemployed cease having any effect

on wage settlements. Another example is based on endogenous human capital formation.

If a boom brings more people into the labor market, the new workers may acquire skills

on the job that naturally augment the supply of labor for the future. Conversely, skills

may atrophy during lengthy spells of unemployment. Hysteresis can also come from

technology shocks, if faster (slower) technological progress is induced by a booming

(slumping) economy.

But theorizing is cheap. The more important question is whether any of these theories

of hysteresis capture the essence of macroeconomic reality. To begin with, does output

actually have a unit root?

Unfortunately, that question is difficult to answer statistically. In a well-known and

provocative paper written nearly two decades ago, Campbell and Mankiw (1987) argued

that it does. But more recent work has emphasized how hard it is to discriminate between

a model with a unit root and a trend stationery model with a root close to but below

unity—especially with relatively short time series. For example, Stock and Watson

(1999, p. 55) estimate that the 90% confidence interval for the largest autoregressive root

in the time series for log real GDP in the U.S. runs from 0.96 to 1.10. For analyzing and

describing very long-run behavior, it makes a world of difference whether the largest root

is, say, 0.98 or 1.00. But, in the short run, these two time-series models are virtually

indistinguishable. Believers in mean reversion have taken solace from this point. But it

also means that it is not easy to dismiss hysteresis.

Assumption 2: The conventional, though much-disputed, effects of fiscal deficits on

interest rates, and thus on the capital stock, leave no lasting imprint on GDP. The old

“crowding-out” argument holds that deficit finance, while expansionary in the short run,

is contractionary in the long run because a larger accumulated public debt leads to higher

real interest rates, and thus to less business investment, and thus to a smaller capital stock

and lower potential output in the future. Indeed, this chain of logic is one of the ideas

behind several of the models that the Congressional Budget Office (2003) recently used

for the “dynamic scoring” of tax cuts. Thus, in several of those models, the estimated

long-run effect of the 2003 Bush tax cut was to reduce real GDP, not to raise it.

Furthermore, in most of what follows, I will adhere to the consensus view by

assuming that:

Assumption 3: Due to some sort of nominal rigidities, real output does respond in the

short-run to aggregate demand shocks, such as monetary and fiscal policy; and:

Assumption 4: The macroeconomy has the natural-rate property, by which I mean (a)

that output returns to potential and (b) that the path of potential output is unaffected by

either monetary or fiscal policy. (However, see Assumption 2 above).

Assumptions 3 and 4 imply that both fiscal and monetary multiplier paths—which,

these days, are often identified with impulse response functions—have characteristic

“hump” shapes, rising to a peak and then falling back to zero. Two examples of such

paths for fiscal policy are displayed in Figure 1. Each requires some explanation, for

standard impulse response functions and multipliers are really two different animals.

Figure 1(a) comes from a VAR analysis of fiscal policy by Blanchard and Perotti

(2002). In interpreting this graph, recall that, in any VAR, a shock sets in motion

complex dynamic reactions that move all the variables. So, in particular, an initial fiscal

shock in the three-variable Blanchard-Perotti model leads to subsequent changes not just

in real GDP, but also in both government spending and taxes. This fact makes

interpreting the impulse response functions in their paper as “fiscal multipliers”

somewhat problematic; and many of them do not have the familiar hump shape. So I

have chosen to display instead the reaction of real GDP to their dummy variable for

1975:2—because that was the quarter of a large, unanticipated income tax rebate. The

dummy, of course, does not change subsequently; and the dynamic responses of

government spending and taxes to this shock shown in Blanchard and Perotti’s paper (but

not repeated here) are very small. So Figure 1(a) comes close to showing the pure effect

of a one-time, non-repeated fiscal stimulus.

Figure 1(b) shows the dynamic multiplier path generated by a simulation of the

Federal Reserve Board’s FRB/US model—a large, structural, macro-econometric model

with forward-looking expectations developed and used by the Fed staff. It simulates the

effect of a sustained rise in government spending, allowing for the partially-offsetting

monetary policy reaction implied by a Taylor rule. But, unlike a VAR, there are no

further (endogenous) responses of other fiscal variables. Thus, although panels (a) and

(b) look similar, the comparison between them is not clean. That said, they both display

the characteristic hump shape.

Blanchard and Perotti (2002, p. 1346) present VAR results with both deterministic and stochastic trends.

Figure 1(a) corresponds to their deterministic trend case.

-2

-4

-6

3 6 9 12 15 18

1.0

0.5

0.0

-0.5

2001 2002 2003 2004 2005 2006

Figure 1(a)

Response of real GDP to a one-shot tax reduction

Figure 1(b)

Response of real GDP to a sustained rise in government purchases

Source: Blanchard and Perotti (2002, Figure IV)

Source: Elmendorf and Reifschneider (2002, Figure 11)

The two panels of Figure 1 should remind us of an important but oft-ignored point about

fiscal policy: “Transitory” does not necessarily mean “fleeting.” In Figure 1(a), the peak

effect on GDP comes after five or six quarters, but there are still notable effects three

years or more later. In Figure 1(b), the peak multiplier (of about 1) is reached

immediately, and then it recedes gradually. But there are still sizable effects one and two

years after the fiscal shock. With real effects lasting as long as these, well-designed

fiscal policy can indeed be used to “fill in troughs and shave off peaks”—which would

appear to make fiscal policy a viable candidate for the role of macroeconomic stabilizer.

And indeed, in the early post-Keynesian period, fiscal policy was expected to play

precisely that role.

So why, then, did educated opinion converge on the proposition that fiscal policy is, to

a first approximation, useless? There appear to be two very different sets of arguments.

Practical/political arguments

The lags depicted in Figure 1 are called outside lags, the time that elapses between a

fiscal policy shock and its effects on the economy. Most evidence suggests that these

outside lags are substantially shorter than the corresponding outside lags for monetary

policy.

But fiscal policy is also subject to potentially long inside lags, the delays between

recognition (by whom?) of the need for fiscal stimulus or restraint and the promulgation

of the appropriate policies. Some of these inside lags occur for compelling administrative

reasons. For example, if Congress decides to stimulate economic activity by building

more public infrastructure, the natural spend-out rate of such programs will probably be

very slow. Speeding up the process artificially for stabilization purposes would be

wasteful. Similarly, when tax changes are made, the Internal Revenue Service needs

some time to change withholding schedules, send out rebate checks, issue new tax forms,

and so on.

Other inside lags occur for political reasons. Even when there is a modicum of

bipartisanship and good will, Congress may need (or take) a long time to reach a decision

on whether and how to change taxes or spending. After all, they don’t call the United

States Senate “the world’s greatest deliberative body” for nothing. Beyond that, political

wrangling can delay Congressional decisions for many months, especially in a

presidential system with weak party discipline like ours, rather than in a more disciplined

parliamentary system like the United Kingdom’s. Delays from this source are particularly

likely when different parties control the White House and Congress. Thus, at least in the

United States, long political lags may be the most cogent argument against discretionary

fiscal policy.

Theoretical/economic arguments

I noted earlier that nominal rigidities are sufficient to imbue aggregate demand shocks

with short-run effects on real output. But some well-known theoretical arguments imply

that fiscal policy cannot even affect aggregate demand. During the long-running

monetarist-Keynesian debate, monetarists argued that fiscal policy was powerless to

move aggregate demand, which was controlled instead by monetary policy--presumably

because the LM curve was vertical. That old debate has a slightly (and deservedly)

archaic ring to it today. For decades, the most-discussed argument for why fiscal policy

might be impotent is the so-called and badly named “Ricardian equivalence”

proposition.

Suppose the pure permanent income hypothesis (PIH) with perfect foresight holds, so

that only present-value budget constraints matter.

Then a bond-financed tax cut simply

defers tax payments until some future time when the interest and principal payments

come due. It does not change the present value of those tax payments because the present

value of the future payments to the bondholders must be exactly equal to the market

value of the bonds, and hence of the tax cut. On the assumption that only present-value

budget constraints matter to consumers, then, spending will be unaffected by what

amounts to a pure shift in timing.

As Barro (1974) and, before him, Patinkin (1956)

pointed out long ago, consumers will simply save their tax cuts so as to be able to make

the interest and principal payments when they come due.

In subsequent sections, I will evaluate both the practical/political and the

theoretical/economic arguments for the futility of fiscal policy. But first, it may be useful

to summarize briefly the interplay of events and ideas that led to such dramatic changes

in the conventional view of the feasibility of conducting a stabilizing fiscal policy. The

roles of both sets of arguments will be apparent.

II. Changing Views: A Brief History of Events and Ideas on Fiscal Policy

The history of thought on fiscal policy since its birth in 1936 divides naturally into

four episodes.

It is badly named because Ricardo did not believe in it. See O’Driscoll (1977).

Perfect foresight is not necessary. Rational expectations will do.

As is well known, this assumes, among other things, that consumers discount future flows at the

government bond rate.

The triumph of Keynesianism, 1936-1966: The first three decades following

publication of The General Theory (1936) were the years of what Herbert Stein (1969)

called The Fiscal Revolution in America. Keynes’ ideas, which emphasized fiscal over

monetary policy,

spread like wildfire. Abba Lerner (1943) wrote of the importance of

using well-timed budgetary changes for “functional finance”—his term for what we now

call fiscal stabilization policy. The early editions of Paul Samuelson’s path-breaking

textbook, Economics (first edition: 1948), explained the use of both fiscal and monetary

policy but clearly emphasized the former. One notable section of that text (pp. 353-354),

entitled “The Inadequacies of Monetary Control of the Business Cycle,” emphasized the

(now) old saw that a central bank can “lead a horse to water but cannot make him drink.”

Mindful of potential political delays, Richard Musgrave (1959) later promoted the idea of

“formula flexibility,” whereby Congress would pre-legislate both the form of and the

trigger (say, a drop in GDP) for future tax or expenditure changes for stabilization

purposes—thereby converting discretionary policy into automatic stabilization.

Others

advocated maintaining a backlog of spending projects “on the shelf,” for use when

cyclical conditions warranted.

The Kennedy-Johnson tax cuts of 1964-1965 marked the first deliberate use of fiscal

policy in U.S. history, and they were judged to be a great success. From a modern

perspective, one can only marvel at the unabashed blend of activism and optimism

exuded by Walter Heller (1966) in his book of reminiscences about the New Frontier.

Heller wrote that both monetary and fiscal policy had “to be put on constant, rather than

intermittent, alert” in order “to provide the essential stability at high levels of

Probably because interest rates were so low during the Great Depression. I will return to this point later.

Seidman (2003) and Solow (2002) have recently tried to revive this idea. Seidman’s book is a

particularly useful reference on many of the points touched upon in this paper.

employment and growth that the market mechanism, left alone, cannot deliver.” To do so,

fiscal policy must become “more activist and bolder,” and “has to rely less on the

automatic stabilizers and more on discretionary action.”

In brief, fine tuning was “in.”

The consensus crumbles, 1967-1977: But it would soon be “out.” A series of adverse

events first shook and then destroyed faith not only in fiscal policy, but in stabilization

policy more broadly. In the space of a scant decade, the old consensus utterly collapsed.

The first blow was the Vietnam War, which piled heavy government spending atop an

economy that was already fully employed. President Lyndon Johnson overrode the

counsel of his Keynesian advisers by insisting on prosecuting the war without either

trimming Great Society spending or raising taxes.

The predictable—and, in fact,

predicted—result was an overheated economy. Soon inflation was on the rise, and

Keynesian economics was being accused, unjustly, of being inherently inflationary.

That charge received apparent support from both the world of ideas and real-world

events. On the intellectual front, Friedman (1968) and Phelps (1968) challenged, and

eventually demolished, the notion that the Phillips curve represented an exploitable long-

run tradeoff. Aiming to keep unemployment below the natural rate, they argued, would

drive inflation ever higher.

On the policy front, what Gordon (1980, p. 136) aptly called “the Waterloo of activist

fiscal stabilization” came when the 1968 tax surcharge failed to curb the Vietnam-

induced inflation.

The failure of the 1968 surtax greatly damaged the idea of using

fiscal policy for stabilization purposes—and in two distinct ways. First, the 2½ year delay

in getting the tax hike enacted illustrated just how painfully long the inside lags in fiscal

Heller (1966). The quotations come from pages 9, 68, 69.

See Okun (1970), especially Chapter 3.

See Eisner (1969) and Okun (1971).

policy could be.

In a world in which recessions typically last less than a year, and an

entire business cycle takes, say, four years, an inside lag of over a year makes fiscal

policy a dubious proposition, at best. Second, Robert Eisner (1969) raised an intellectual

conundrum. Activist use of tax policy for stabilization purposes would seem to call for

temporary, and perhaps even frequent, changes in income taxes. But the PIH implies that

such tax changes, if believed to be temporary, should have only small effects on

consumer spending.

So repetitive use of the income-tax weapon for stabilization

purposes should severely undermine its efficacy.

These dual failures seemed to be replicated in the deep recession of 1974-1975 when

first President Nixon and then President Ford failed to recommend anti-recessionary

policies until it was too late.

Then the temporary nature of the 1975 tax cut undermined

its effectiveness. My study of the 1968 and 1975 episodes together (Blinder [1981])

concluded that the two temporary taxes had about half as much short-run impact on

aggregate demand as equal-sized permanent tax changes would have had.

Long inside lags, weak tax effects due to the PIH, and the vertical long-run Phillips

curve have precious little to do with the monetarist claim of fiscal impotence owing to a

vertical LM curve. But all these problems with fiscal policy seemed to get mixed up

together in the anti-Keynesian backlash, and fiscal stabilization fell deeply out of favor.

Its nadir may have come when President Carter’s call for a short-term fiscal stimulus in

1977 was swiftly rejected by Congress—an event that would be repeated 16 years later

for President Clinton.

Johnson’s advisers urged a tax hike on him as early as late 1965 (see Okun (1970)). LBJ resisted until the

middle of 1967, when he recommended a temporary income tax surcharge. Then Congress then took about

18 months to enact one.

This very point resurfaced recently in the context of the Bush tax cuts in 2001.

But Congress acted speedily this time, demonstrating that the inside lag could be short.

Huge deficits crowd out stabilization policy, 1981-2001: President Reagan’s massive

tax cuts proved to be another landmark in the history of fiscal policy. Despite the

Reaganite attack on the weak Carter economy, the 1981 tax cuts were justified not by

Keynesian aggregate demand considerations—which were denigrated—but by a new

doctrine called supply-side economics. This is not the place to discuss that ill-fated (and,

some would say, silly) doctrine, other than to observe that it helped pave the way for a

huge multi-year tax cut which ushered in an era of chronically-large federal budget

deficits.

The Reagan legacy of huge deficits “as far as the eye can see” fostered a dramatic

repositioning of fiscal policy—away from (cyclical) stabilization policy and toward

(secular) deficit reduction. The new-found devotion to fiscal prudence grew to be so

extreme that, in 1985, Congress passed the Gramm-Rudman-Hollings Act which, had it

actually been followed, would have short-circuited even the automatic stabilizers by

requiring strict adherence to annual targets for the federal budget deficit (an endogenous

variable). Five years later, when the economy slipped into recession, fiscal stimulus was

considered out of the question. Instead, taxes were increased as part of the 1990 deficit-

reduction package. By then, even Keynesian economists were so desperate for deficit

reduction that they accepted this procyclical tax hike without protest.

After the election of 1992, things went quite a bit further. President Clinton’s original

budget proposal combined substantial long-run fiscal consolidation with a small, short-

run fiscal stimulus—a strategy of one step backward, five steps forward. But this two-

pronged strategy proved to be too clever by half, and the stimulus part was quickly

rejected by Congress. Instead, a deficit-reduction package even larger than the one

Clinton had proposed (barely) passed. So Clintonomics tuned out to be about, first,

reducing the deficit, then balancing the budget, and finally building a sizable budget

surplus. (There are worse policies, and we have them.)

The new political realities of the 1980s and 1990s were reflected rapidly in academic

thinking. Scores of papers appeared on the effects (or lack thereof) and the sustainability

(or lack thereof) of budget deficits. Tellingly, the 1986 NBER conference volume, The

American Business Cycle: Continuity and Change (Gordon [1986]), did not even include

a chapter on fiscal policy—despite the cyclical focus of its title. Instead, there was a long

essay by Barro (1986) on “The Behavior of United States Deficits,” which focused on the

tax-smoothing hypothesis (Barro [1979]).

The fact that the Clinton boom started almost immediately after Congress passed a

budget reduction package gave rise to some rethinking—some of it serious, some of it

muddled—of even the sign of the fiscal-policy multiplier. Among politicians and media

types, the notion that raising taxes and/or cutting spending would expand (rather than

contract) the economy took hold rapidly and uncritically—with seemingly little thought

about exactly how this was supposed to happen. Quicker than you can say “Robert

Rubin,” the idea that reducing the budget deficit (or increasing the surplus) is the way to

“grow” the U.S. economy—even in the short run—came to dominate thinking in

Washington. This thinking was, of course, profoundly anti-Keynesian.

In the academic world, some earlier theoretical research by Turnovsky and Miller

(1984) and Blanchard (1984) was dusted off and used to explain how a credible reduction

in expected future budget deficits could in fact increase aggregate demand by lowering

long-term interest rates today. Those models, of course, did not claim that a reduction in

the current budget deficit would be expansionary. Still, the Turnovsky-Miller-Blanchard

thesis offered one theoretically coherent explanation of the Clinton boom.

But could

the lessons of those glory years be generalized? Few people asked the question.

The new era, 2001-?? : It is hard to know how to characterize the fiscal policy of the

current President Bush. The ideas that eventually morphed into the tax cuts of 2001-2003

began as campaign promises in 1999. Given what has happened since, the original

argument sounds like a bad joke: The federal government should “give back” the money

to the people rather than run excessive budget surpluses. The tax cuts were most

emphatically not recommended for short-run stabilization-policy purposes.

In fact, the

Federal Reserve was worried at the time that the U.S. economy might be overheating. But

when the economy slowed in 2000, and then sagged in 2001, the Bush administration

quickly changed its rationale for the tax cuts to the more traditional Keynesian one: The

economy needed stimulus. But the policy itself barely changed.

In terms of our brief history of events and ideas, three remarkable things have already

happened during the presidency of George W. Bush. First, consistency proved to be the

hobgoblin of small minds. Without skipping a beat, both political parties and most of the

press jettisoned the Clinton-era view that deficit reduction was the way to stimulate the

economy and returned to the older Keynesian notion that deficit expansion would do the

trick—apparently, without noticing the inconsistency.

Second, a political consensus in favor of fiscal stimulus formed quickly and decisively

in 2001—so quickly that both the 2001 and 2003 tax cuts were enacted in a matter of

There were also many incoherent ones, usually involving the restoration of confidence.

For more on this subject, see Blinder and Yellen (2001, Chapter 4.)

Nor was it ever argued, even by opponents, that lower taxes that led to a deficit would slow down

economic growth!

months, thereby demonstrating that the inside lags in fiscal policy could really be quite

short, even in a narrowly-divided Congress.

Third, yet another old Keynesian idea—the liquidity trap—rose like Lazarus from the

tomb of discarded notions.

As the Fed lowered the federal funds rate toward 1%, and

the economy still did not revive, economists began to express concern about the zero

bound on nominal interest rates—a trap that had already ensnared the Bank of Japan and,

on that account, had engaged the interest of a number of academic and Federal Reserve

economists.

Briefly, the problem is this. Once the central bank lowers the overnight interest rate to

zero, it is left with only “unconventional” monetary tools because base money and short-

term debt become perfect substitutes. But these unconventional weapons are weaker than

the conventional weapon: lowering the short-term interest rate by purchasing short-term

government paper.

For example, the central bank can always expand the monetary base by purchasing

long-term bonds rather than short-term issues. But that is equivalent to a two-part policy

in which the central bank first purchases short-term debt in the open market (thereby

creating bank reserves) and then turns around and sells this debt to purchase an

equivalent amount (at market value) of long-term debt. The first part of this operation is

a conventional open-market purchase, which has no effect when the nominal interest rate

is zero. The second part is a dose of “Operation Twist” which, we have been conditioned

to believe, does not accomplish much. If that is what monetary policy comes down to at

zero nominal interest rates, then fiscal policy, for all its flaws, starts to look like a viable

See Krugman (1998).

For example, it was the subject of a Federal Reserve conference in October 1999. See the November

2000 special issue of the Journal of Money, Credit and Banking.

option after all. And, indeed, some of the most compelling suggestions for ending

Japan’s deflationary slump combine expansionary monetary and fiscal policy in

monetized deficit spending (or tax cutting).

Were it not for the fact that the ink is not yet dry on this fourth episode in U.S. fiscal-

policy history, it would be tempting to say that we have come full circle. Just think about

what has been restored since George W. Bush took office: belief in fiscal stimulus, belief

that the inside lags in fiscal policy are short, and skepticism about the efficacy of

monetary policy. This is a world that Walter Heller would recognize.

With this as background, I now turn to some of the economic/theoretical arguments

that have been made against the efficacy of fiscal stabilization policy.

III. Temporary Income Taxes and Present-Value Budget Constraints

Theory

The structure of the basic argument against the use of temporary income tax changes

as stabilization devices is very straightforward—much simpler than much of the

literature, with its emphasis on intergenerational transfers, bequest motives, transversality

conditions, and the like. Here is the simplest non-stochastic version of the Ricardian-

equivalence argument.

Suppose a representative consumer’s current spending depends only on the present

discounted value of her lifetime resources:

= A

+ Σ

t+i

Among the many sources that could be cited, see Bernanke (2000).

where A

is current net worth, y

t+i

is future after-tax earnings, and δ

is the appropriate

discount factor for cash flows at date t+i. This is the central assumption; as Hillel said

(though in an admittedly different context), all the rest is commentary.

Now consider a tax cut of ∆y financed by issuing bonds. Current receipts rise to y

∆y. But future taxes must rise by just enough to meet the interest and principal payments

on the bonds—meaning that the present value of these future taxes must rise by exactly

∆y. Thus W

is not changed by this fiscal operation, which alters only the timing of

receipts and not its discounted present value. In consequence, consumption is also

unchanged.

This basic argument can be, and has been, gussied up in many ways. But, in essence,

all the fancier variants come down to the same simple argument I just made. What can go

wrong with this argument? Many things. But since most of the objections to Ricardian

equivalence are so familiar, I will list them very briefly.

1. Bequests: Suppose some of the future tax burden falls on generations yet unborn.

That part cannot, of course, affect their spending today; they are not alive yet. But Barro

(1974) pointed out long ago that, under one particular specification of intergenerational

altruism, today’s consumers will essentially act as farsighted agents for their heirs—and

adjust their bequests so as to make debt and taxes equivalent. Of course, Barro’s model is

not the only possible model of the bequest motive, and much ink has been spilled over

this issue. But, in my view, most of the debate is beside the point because, in the real

world, the bonds that will be issued to cover deficits will almost always mature in less

than 10 years—a time frame within which most of today’s taxpayers will still be around

to pay the bills. So intergenerational aspects of present-value budget constraints are

mostly irrelevant.

2. Liquidity constraints: Current consumption may not depend only, or even mostly,

on the present-value budget constraint. If liquidity constraints are binding, for example,

current income will matter more than future income because it loosens liquidity

constraints. In that case, a debt-financed tax cut will raise spending. Even if only a

portion of the population is liquidity constrained, as the evidence suggests, Ricardian

equivalence will fail. Because they are so important to the central questions of this paper,

I will return to liquidity constraints later.

3. Different discount rates: The simple present-value argument assumes that taxpayers

and bondholders discount future cash flows at the same rate. But if taxpayers discount the

future at an interest rate higher than the government bond rate, the present value of the

current purchasing power gained will exceed the present value of the future purchasing

power lost—and consumption will rise in consequence. A variant on this objection is:

4. Myopia: Homo sapiens may not be as farsighted as homo economicus. Real people,

it appears, give insufficient weight to the future or, what comes to the same thing,

discount future flows at extraordinarily high rates or have short planning horizons. For

such (real) people, the rise in current income is a stronger influence on current

consumption than the fall in future income.

5. Precautionary saving: Precisely this last sort of behavior can even be rationalized

(for many reasonable utility functions) on optimizing grounds by the theory of

precautionary saving.

Receiving more income today and expecting to receive less

income in the future reduces income uncertainty, which reduces the need for

See, for example, Barsky, Mankiw, and Zeldes (1986).

precautionary saving. As long tax payments rise with income, such a swap of present for

future income can lead to higher spending today.

6. Consumer spending may react more than consumption: Current tax receipts that are

not spent must be saved. One way to “save,” by economists’ definition, is to purchase a

consumer durable that yields a flow of consumption services into the future. But that part

of saving actually adds to current aggregate demand.

While each of the six objections summarized above is familiar, a seventh one seems

not to be:

7. The present-value government budget constraint is irrelevant in practice. Modern

economic models lean heavily on the so-called present-value government budget

constraint (PV-GBC)—which ensures, for example, that any additional deficit run today

must be balanced by surpluses eventually because the debt cannot explode upward

forever. It is the PV-GBC that makes every income tax change, in a sense, temporary.

Let’s grant this point. But note that the transversality condition from which the PV-GBC

is derived holds only asymptotically.

As Ronald Reagan proved in the 1980s, and as

George W. Bush may be proving again today, the government budget can traverse an

explosive debt path for a decade or two without any cataclysmic consequences. The PV-

GBC is thus a theoretical nicety that places no meaningful constraint on policy today,

next year, or for, say, the next decade or two.

Evidence

Okun (1971) was the first to study the temporary tax issue empirically. His context

was the explicitly temporary income tax surcharge enacted in 1968. Using the

For example, Canzoneri, Cumby, and Diba (2002, footnotes 22 and 23) note that the PV-GBC is derived

from the government’s flow budget constraint and the transversality condition of the household’s

maximization problem.

consumption equations of four different econometric models, he compared the “full

effect” view (that the surcharge reduced spending by as much as a permanent tax increase

would have) to the “zero effect” view (that the surcharge did not affect spending at all).

Okun concluded that the “full effect” view explained the data better. However, Solow

and I (1974, pp. 107-109) subsequently showed that an intermediate “50% effect view”

fit Okun’s data better than either extreme.

Modigliani and Steindel (1977) and Eckstein (1978) conducted similar studies of the

effects of the 1975 tax rebate, using the consumption equations of two large-scale

econometric models. Like Okun, each found sizable effects on spending—although

Modigliani and Steindel (1977) expressed skepticism about the model’s predictions.

Some years later, I re-examined the 1968 and 1975 episodes together (Blinder [1981]),

using a more complex specification that treated a temporary tax change as a weighted

average of a permanent tax change and a pure windfall. (Note that a windfall does not

have “zero effect” on spending under the PIH.) The point estimate of the weighting

parameter was exactly 0.50, although its standard error was a large 0.32. However, when

Deaton and I (1985) re-examined this question with a different model and, perhaps more

important, revised data, we found something closer to Okun’s original “zero effect” view.

This older time series literature tested the PIH by asking whether consumers’

responses to explicitly temporary income changes are larger than predicted by theory.

But, as Deaton and I (1985, p. 498) concluded, the time series data offer so few

observations on temporary taxes that the “results are probably not precise enough to

persuade anyone to abandon strongly held a priori views.” A newer strand of research,

influenced profoundly by Hall’s (1978) rational expectations approach to the

consumption function, poses a different, though related, question: Is the response of

consumers to easily-predictable income changes greater than theory suggests? And it

seeks answers mainly in cross-sectional data.

Some of this research takes advantage of what might be called natural experiments. It

began with a clever paper by Shapiro and Slemrod (1995), who noted that the first

President Bush conducted a rather curious tax experiment in 1992: He reduced

withholding rates by executive order beginning in March of that year even though

Congress had not cut income tax rates! Thus American taxpayers were treated to higher

cash flow (but no higher accrued after-tax income) over the last ten months of 1992, only

to owe more in taxes in their April 15, 1993 settlements. The Bush experiment therefore

amounted to a very temporary increase in disposable income that was quickly reversed.

According to the PIH, consumers should have ignored such a change in the timing of

receipts, exactly as in the Barro (1974) model.

Yet nearly half of the respondents told

the University of Michigan’s survey takers that they would spend “most” of their (very

temporary) increase in take-home pay. A similar subsequent study by the same authors

(Shapiro and Slemrod [2003]) of the so-called income tax “rebate” of 2001 found that

only 22% of respondents said they would spend “most” of it.

Parker (1999) exploited the fact that, for a minority of workers each year, the payroll

tax for Social Security falls abruptly to zero when their earnings rise above the Social

Security maximum, and then suddenly jumps back to normal again on the following

January 1st. Under the pure PIH, such predictable, seasonal fluctuations in after-tax

Well, not quite exactly. Since the government gave taxpayers interest-free loans, the present value of

lifetime resources was raised slightly.

income should have no effect on spending. But Parker (1999) found that they do; in fact,

he estimated a marginal propensity to consume of about 0.5 over a three-month period.

Similarly, when Souleles (1999) studied consumer responses to income tax refunds—

another predictable source of after-tax income—he found an MPC of around 0.6. And, in

a subsequent study, Souleles (2002) found that, even though the phased-in Reagan tax

cuts were pre-announced and predictable, people did not spend their additional after-tax

income until they had the money in hand. The estimated MPC for non-durables was 0.6

or greater when the taxes were actually cut. This last finding echoes what Deaton and I

(1985) had found years earlier in studying consumer responses to the Reagan tax cuts.

Hsieh (2003) reported some puzzling findings for Alaskan families. Their spending

seemed not to react to the relatively large and predictable annual payments from the

Alaskan Permanent Fund (which come from oil revenues), but did react strongly to

relatively small and predictable income tax refunds—just as Souleles (1999) had found.

Most recently, Johnson, Parker, and Souleles (2004) assessed the spending effects of

the so-called 2001 tax rebate. This episode was interesting for two reasons. First of all,

while widely described as a “rebate,” the 2001 tax cut was actually an early installment

payment on a permanent tax-rate reduction. Second, for administrative reasons, the

checks were sent out on a randomized basis. This latter feature enabled Johnson et al. to

estimate sizable initial-quarter spending responses with some precision.

While these two strands of the consumption literature ask different questions, their

respective answers have a certain consistency: Both point strongly toward the importance

of binding liquidity constraints. In the time-series literature, it is presumably liquidity

constraints that make consumers react much more strongly to current cash income than

the PIH says they should. In the cross-section literature, liquidity constraints are the

presumptive reason why households react much less strongly to anticipated future

income—even when the “future” is not very far off. Instead, they wait until they have

their hands on the money, just as the time series evidence suggests.

The lesson for stabilization policy, therefore, seems clear: Even temporary income tax

changes can pack substantial punch, though perhaps not quite as much as a permanent tax

cut. Deaton (1992, pp. 101-102) had it about right a dozen years ago, when he wrote in

his survey of consumption that “if macroeconomic policy-makers wish to use taxes to

fine-tune the economy,… then the empirical failures of the [permanent income] theory

[with rational expectations] are certainly large enough to make a big difference.” Perhaps

we economists have taken the PIH too much to heart.

IV. Temporary Tax Changes and Intertemporal Substitution

Making an income tax change temporary probably undermines its effectiveness, at

least somewhat. But there are other sorts of tax changes which become more powerful

when they are made temporary. I refer, of course, to taxes that create incentives for

intertemporal substitution, such as investment tax credits, value-added taxes, sales and

excise taxes, and the like.

The idea is simple enough. Consider a one-year reduction in a consumption tax from

to τ

, reverting back to τ

next year. The relative price of goods next year versus

goods this year will rise from P

t+1

/(1+r

≡ λ

when the tax rate is the same in both

periods to:

t+1

(1 + τ

)/(1+r

(1+ τ

) = [(1 + τ

)/(1+ τ

)]λ

> λ

when the tax rate is τ

this year and τ

next. In theory, this change in relative prices

should redirect spending from next year to this year.

Can intertemporal tax incentives like this be used effectively as instruments of fiscal

policy? Sumner (1979) found little evidence that temporary changes in Ontario’s retail

sales tax had an extra-large impact on consumer spending due to intertemporal

substitution.

In the United States, there is no value-added tax, and sales taxes are the province of

the states. So the main intertemporal tax policy that has actually been utilized as part of

U.S. fiscal policy is the investment tax credit. The credit was invented by President

Kennedy’s Council of Economic Advisers and introduced in 1962—for Keynesian

reasons by the way—at a 7% rate. Between then and its abolition as part of the Tax

Reform Act of 1986, the ITC was suspended or repealed twice and had its rates

readjusted twice—often for cyclical reasons.

ITCs have also been implemented in a

number of other industrialized countries, often under a different name.

Economic theory strongly suggests that the credit should be more powerful when it is

enacted on a temporary basis. In his famous paper on econometric policy evaluation,

Lucas (1976, p. 30) observed that: “The whole point, after all, of the investment tax credit

is that it be viewed as temporary, so that it can serve as an inducement to firms to

reschedule their investment projects.” Yet when the ITC was made “permanent” in 1979,

two years after what I labeled the “nadir” of fiscal policy, the U.S. Treasury (1979, p.

365) stated emphatically that “changes in the investment tax credit rate should not be

considered in terms of short-run stabilization objectives.”

See Chirinko (1999), which is a particularly useful source of information on the ITC.

See, for example, Jorgenson and Landau (1993).

Econometric appraisals of the “bang for the buck” effectiveness of the ITC have given

the credit mediocre reviews.

One reason may be that the credit has never been made

marginal—it has always been applied to all qualified investments. Economists in the

1992 Clinton campaign had persuaded the candidate to propose a marginal ITC as a low-

cost way to provide some fiscal stimulus. But the idea was quickly scrapped in 1993 for

lack of support in Congress. I still think the idea is a good one.

The 2002 tax cut bill included a provision that offered accelerated (called “bonus”)

depreciation for about 18 months. A year later, the “bonus” was increased and the period

slightly extended (to the end of 2004) as part of the 2003 tax cut. The idea, of course,

was exactly the same as that behind the ITC: to put investment goods “on sale” for a

while, and thereby to encourage intertemporal substitution. Nonetheless, a controversy

arose at the time, with economists in the Bush administration claiming that the bonus

depreciation provision would be more powerful if enacted for a longer period! It is, of

course, too early for there to have been any econometric studies of this most recent

episode.

V. Countercyclical Variations in Government Purchases

As mentioned earlier, one stabilization-policy idea that dates back to the early

Keynesian period is the use of timely variations in expenditures on “public works” to

smooth cyclical fluctuations. While the roots of this idea are thoroughly practical and

atheoretical, it makes good theoretical sense on allocative grounds—at least in principle.

After all, periods of slack resource utilization are times in which the shadow values of

See Auerbach and Hassett (1991), and Chirinko (1993) for a survey.

factor inputs are presumably low in the private sector. What better time to put those

resources to use for public purposes?

Barro (1981) found that federal government defense purchases have a significant

positive impact on real output, with temporary changes (mainly associated with wars)

having larger effects than permanent ones. He argued that this finding provides support

for a theoretical model in which temporarily higher government purchases raise the real

rate of return, thereby inducing intertemporal substitution in both consumption (less

today, more tomorrow) and labor supply (more today, less tomorrow). But Barro’s

theoretical rationale (intertemporal substitution) and his empirical results (wars boost real

output) are sufficiently disconnected that one can accept the latter without buying into the

former.

The major objections to using public expenditures as a countercyclical weapon seem

to be more practical than theoretical. But I think they are powerful nonetheless. To begin

with, wars do not seem like particularly promising devices for stabilization policy!

More seriously, there are normally quite lengthy lags in the political process before

new spending projects are authorized by Congress. Then, since authorizing committees

and appropriating committees are different, still more time elapses between legal

authorization and the actual appropriation of funds. These legislative lags could

conceivably be short-circuited by having a queue of projects pre-authorized, pre-

appropriated, and sitting “on the shelf” ready to go when the cyclical need arose. But I,

for one, have a hard time imagining the U.S. Congress doing anything like that. And

even if the lags in the authorizing and appropriating processes could be completely

eliminated, the slow natural spend-out rates of most public infrastructure projects remains

a serious handicap. For example, out of each $1 appropriated for highway expenditures,

less than one-third is likely to be spent within a year. Accelerating the pace of spending

on public works for stabilization purposes would be inefficient and wasteful.

To my mind, this all adds up to a recognition that the inside lags for many sorts of

government purchases are lengthy enough to vitiate their usefulness for stabilization

policy. The idea works in theory, but not in practice. If fiscal policy is to be used for

stabilization purposes, taxes (and transfers) are probably the instrument of choice.

VI. Is There a Case for Streamlining Fiscal Policy Institutions?

This discussion points to the long inside lags as perhaps the most critical element of

the case against discretionary fiscal policy. But these lags are not immutable. The

sources of many, if not most, of them lie in policymaking institutions that can be

changed—at least in principle. And a number of suggestions for doing just that have been

made over the years.

One such idea, formula flexibility in setting income-tax rates or public expenditures,

was discussed earlier. Its main virtue is both obvious and substantial. If what we now

think of as discretionary policy changes for stabilization purposes could somehow be

made automatic, then the lengthy inside lags in fiscal policy could be reduced

dramatically. Since the outside lag for most garden-variety fiscal policy changes are

relatively short, the feasibility of conducting a stabilizing fiscal policy would thereby be

greatly enhanced.

What’s the down side? For (good) reasons elucidated earlier, most discussions of

formulaic fiscal responses have focused on taxes rather than government spending. But,

as noted above, temporary changes in income tax rates are believed to elicit muted

spending responses. Perhaps more important, Congress has not shown the slightest

inclination to relinquish any of its ability to bestow gifts upon taxpayers when the

economy is weak. And symmetry does not rescue us when the economy is strong. It is

true that members of Congress wish to avoid the blame for raising taxes at times of peak

demand. But they have a straightforward way to accomplish that objective right now: In

a reversal of the old Nancy Reagan motto, they just don’t do it. The last time Congress

enacted a tax increase aimed squarely at reducing aggregate demand for stabilization

purposes was in 1968.

Instead, it lets the Fed do all the dirty work by raising interest

rates. So it is not at all surprising that Congress has shown no interest whatsoever in

formula flexibility.

A related ivory-tower idea should be mentioned in this context. In a paper published

seven years ago (Blinder [1997]), I asked why some economic decisions are delegated to

unelected technocrats while others are reserved for politicians. One important specific

example of this question is: Why do just about all countries put monetary policy in the

hands of independent central bankers, and yet leave tax policy in the hands of elected

politicians? I went on to speculate on whether technocratic decisionmaking on tax policy

might produce better outcomes than political decisionmaking, suggesting that the answer

might indeed be yes.

In broaching the idea of transferring some aspects of tax policy from the political

sphere to the technocratic sphere, I was not thinking about stabilization policy, but rather

about getting the details of the tax code—with their complex allocative and distributive

Congress did raise taxes in 1983, 1990, and 1993. But in none of those cases was cyclical restraint the

main reason

effects—right. However, the same point applies to getting the timing right in a business-

cycle setting, as advocates of formula flexibility realized many decades ago.

Suppose a group of technocrats, modeled on the Federal Reserve Board, were

empowered to make decisions on the level of taxation, subject to (potentially numerous)

constraints laid down by Congress. Under that institutional structure, the possibility of

conducting a timely and rational fiscal policy would be greatly enhanced. Of course, the

probability that Congress would delegate such authority is probably roughly equal to the

probability that the Red Sox will win the World Series. But fans can dream.

Subsequently, the Business Council of Australia (1999) picked up on this idea and

advocated the establishment of an independent fiscal-policy agency for Australia, along

the lines just suggested. In one version of their proposal, a new agency as independent as

the Reserve Bank of Australia would actually be given the power to make small, across-

the-board adjustments in personal and/or corporate tax rates for stabilization purposes—

unless their order was publicly and explicitly countermanded by the government. In a

softer version of their proposal, the new agency would be merely advisory, making public

recommendations to the government.

A series of related proposals has been made for the euro zone, although the focus there

has been on (secular) budget discipline rather than on (cyclical) stabilization. The much-

maligned—and, one might say, much-ignored—Stability and Growth Pact (SGP) requires

member governments to limit budget deficits to no more than 3% of GDP. Even before

the pact was agreed upon, critics noted that the 3% limit could in principle interfere with

the workings of the automatic stabilizers because it was phrased in terms of actual budget

deficits rather than cyclically-adjusted deficits. So if weak economic performance

lowered tax revenue and raised social welfare expenditures sufficiently, even a

“responsible” fiscal policy could produce a deficit in excess of the 3% limit—thereby

requiring offsetting fiscal actions that are procyclical.

In practice, European

governments have shown themselves unwilling to take such actions, preferring to violate

the pact instead. So the pact has become something of an embarrassment.

Notice the analogy to the old formula flexibility discussions in the United States. In

principle, the SGP requires discretionary responses to (certain) changes in economic

activity. But those changes have proven difficult or impossible to sustain politically, and

they may not make good economic sense, anyway (e.g., if they are procyclical). So some

economists have proposed institutional changes that would make long-run fiscal

discipline somewhat closer to automatic, while still allowing for cyclical responses. For

example, von Hagen and Harden (1995), Wyplosz (2002) and others have called for

replacing the SGP’s excessive deficit procedure with a council of experts not unlike the

“softer” version of the Australian proposal.

This group of technocrats would report and

opine—quite publicly—on the sustainability of the fiscal programs of the euro-zone

governments. The idea would be to bring public and market pressure to bear on

governments that insist on pursuing unsustainable policies.

VII. Out of the Detritus: Some Creative Ideas for Fiscal Stabilization

A short summary of the conclusions so far might run something like this. The

theoretical arguments against the efficacy of fiscal policy as a stabilization tool turn out

Qualitatively, the analogy to the central problem with the Gramm-Rudman-Hollings Act in the United

States is almost perfect. However, marginal tax-and-transfer rates are much higher in the euro zone than in

the United States, making the quantitative dimensions of the problem more severe in Europe. (See

Canzoneri, Cumby, and Diba (2002), pp. 340-343.)

For a summary of and rationale for such proposals, see Fatas et al. (2003).

to be pretty thin gruel. But the practical arguments seem much more compelling.

Timely variations in government purchases (say, public works) for stabilization policy,

though fine in theory, do not appear to be either sensible or workable. Changes in taxes

and/or transfer programs are far more suitable for stabilization purposes, but current

institutional arrangements make prospects for success slim. Nor do any of the

institutional changes that would make successful fiscal stabilization more achievable

seem likely to be adopted. So maybe the current conventional wisdom is right after all.

Before giving up, however, let us consider some creative suggestions for stabilizing

fiscal policies that have been made in recent years. Each of them is designed to address

the main perceived weakness of using temporary income tax changes to alter consumer

spending: Elementary theory says that temporary changes in income taxes should yield

less aggregate-demand “bang” for each income-loss “buck” than permanent tax changes.

Yet the rhythm of the business cycle virtually dictates that tax-transfer changes for

stabilization purposes should be temporary. What to do?

Targeting tax-transfer payments better

One response suggested by the empirical literature would be to target tax-transfer

changes made for stabilization purposes on those people who are most likely to be

liquidity constrained, and therefore to have MPCs at or near one. To some extent, that

means targeting income tax changes on lower-income households, who are more likely to

be living hand to mouth. There are two drawbacks.

First, the suggested remedy is strikingly asymmetric. When the economy needs

stimulus, targeting income tax reductions and increases in transfer payments

disproportionately on the poor serves both stabilization and distributional objectives

admirably. But the idea of targeting income tax hikes or cuts in transfers on the poor

when the economy needs restraint is repugnant to most people. On the other hand, as

noted earlier, Congress never uses fiscal policy to “shave off peaks” anyway. If

discretionary fiscal policy is only used when stimulus is called for, maybe this problem is

not important in practice.

Second, income is an imperfect predictor of who is and who is not liquidity

constrained; the ratio of assets to income may make more sense on theoretical grounds.

In principle, large negative transitory income should be a better predictor of who is

constrained, since a sizable negative income shock should indicate a strong likelihood of

a binding liquidity constraint. And, since transitory income and current income are

highly correlated,

current income may be a decent statistical proxy. Nonetheless, it

would be nice if we could do better. Two suggestions have been made in that regard.

One idea is to use receipt of unemployment insurance (UI) benefits as a proxy for

being liquidity constrained. After all, most people who are collecting UI have recently

suffered a severe drop in earnings (about 50% on average), making their transitory

incomes negative and large. If these people are striving to maintain their previous

consumption levels, they are probably liquidity constrained.

Extending UI benefits during times of high unemployment has indeed become almost

standard practice in the United States.

An additional 13 weeks of coverage, beyond the

usual 26 weeks, is triggered automatically in a particular state once the level of insured

See, for example, Zeldes (1989). But Jappelli (1990) finds that people with lower income are indeed

more likely to be liquidity constrained. Age and wealth are also relevant, however.

For example, Hall and Mishkin (1982) estimate that the variance of the innovation to transitory income is

more than twice as large as the variance to the innovation of permanent income.

However, the Bush administration and the Republican-controlled Congress refused to do so in late 2003,

despite objections from Democrats.

unemployment breaches certain levels.

And Congress often enacts additional

discretionary increases in UI coverage that become effective during, and especially after,

recessions. Data on payments under the Extended Benefits program display sharp spikes

in 1976, 1983, 1992-1993, and 2002-2003

—all years following recessions.

During the Congressional debate over the 2001 stimulus package, a number of

Democrats and liberal economists argued that UI benefits should be extended in time and

broadened in coverage—e.g., by making part-time workers eligible.

The main problem

with this idea appears to be magnitudes. Feasible policy changes in UI benefits are

simply not big enough to combat a recession. Or are they? Let’s look at some numbers.

Between the years 2000 (when the unemployment rate averaged 4.0%) and 2002

(when the unemployment rate averaged 5.8%), total UI benefits increased from $20.5

billion to $42.1 billion.

Let’s imagine that aggressive policy changes might have

boosted that $21.6 billion increase by 50%—that is, by another $10.8 billion—which

seems a high-end estimate of what Congress might actually have done. Assuming an

MPC of one, which is probably too high, and no multiplier effects, which is probably too

low, those additional UI payments would have raised GDP by $10.8 billion. By

comparison, the peak-to-trough decline in real GDP actually experienced between 2000:4

and 2001:3 (expressed in 2002 dollars) was $55.2 billion. And the 2001 recession was

very mild by historical standards. So clearly, no conceivable expansion of UI benefits

can make a big dent in a large recession.

Some states add an additional seven weeks.

See http://workforcesecurity.doleta.gov/unemploy/content/chartbook/images/chtb1.gif on the Department

of Labor website.

See, for example, Krueger (2001).

Economic Report of the President, 2004, Table B-45, p. 337.

Nonetheless, in designing stabilization policies, we should be thinking about

mitigating recessions, not eliminating them. And in that context, discretionary variations

in UI benefits may deserve a more prominent role than they have been given to date. By

like reasoning, a more generous UI program would be a better automatic stabilizer.

A second idea along these same lines, which was suggested by several Democratic

politicians in 2001, is temporarily rebating the “first part” of the payroll tax. Here the

numbers are potentially much larger. The Social Security Administration reports 144.8

million wage and salary workers in 2002. Of these, 103.5 million workers earned

$10,000 or more, and hence would have been eligible for the full $620 tax cut—for a

total expenditure of $64.2 billion. A rough estimate of the value of the 6.2% tax cut to

the 41.3 million workers with covered earnings below $10,000 adds another $10.7

billion

—raising the total cost to about $75 billion. That is more than enough to “fill in

a trough.” The problem in this case is targeting.

Under this particular payroll tax rebate

plan, even middle- and upper-income households will receive temporary tax cuts. Many

of them—perhaps the majority, weighted by income—will not be liquidity constrained.

Exploiting Intertemporal Substitution

Two other recent fiscal policy suggestions seek to exploit the idea that, unlike income

taxes, variations in sales taxes are likely to be made more powerful by enacting them on

See http://www.ssa.gov/policy/docs/statcomps/supplement/2003/4b.html#table4.b7.

That source divides the 41.3 million sub-$10,000 workers into three ranges:

Income range Workers (millions)

0-$999 8.24

$1,000-$4,999 17.62

$5,000-$9,999 15.44

The calculation in the text assumes that the average earnings in each of these three brackets is the midpoint

of the bracket.

Another problem is that the Social Security Trust Fund cannot spare the revenue. But this problem can

be overcome by using general revenue to replace any payroll-tax receipts that the Trust Fund loses.

an explicitly temporary basis.

As a way to bring incentives for intertemporal substitution to bear on stimulating

consumer spending, Martin Feldstein (2001) suggested temporarily suspending Japan’s

5% value-added tax (VAT), and following that by an increase two years later. He

subsequently offered a more complicated version of this idea (Feldstein (2002)): The

government of Japan should embark on a multi-year plan of simultaneously raising the

consumption tax and reducing the income tax, in a balanced-budget way. The idea is the

same in each case: to create incentives for consumers to buy now rather than in the

future.

That was the same idea behind a suggestion I made during the debate over the 2001

stimulus package in the United States (Blinder (2001)). The federal government in the

U.S. has neither a VAT nor a general sales tax. But 45 of the 50 states have the latter. So

I suggested that the federal government offer to replace the lost tax revenue of any state

that would agree to cut its sales tax (up to some limit) for the next twelve months. Of

course, I would not want to exaggerate the impact of such a policy, given the low rates of

sales taxation in the United States and the modest degree of intertemporal substitution

suggested by econometric studies. A temporary and marginal ITC might be a more

potent option.

VIII. Wrapping Up: Is There Anything New Under the Sun?

Today’s conventional wisdom holds that discretionary changes in fiscal policy are

unlikely to do much good, and might even do harm. Why is that? Because the lags in

fiscal policy, especially the inside lags, are long—perhaps longer than the duration of a

typical recession. Because the most plausible fiscal policy weapon, changes in personal

income taxes (or transfer payments), is likely to be weakened by deploying it on a

temporary basis. And because an obviously superior stabilization weapon—namely,

monetary policy—is readily available.

When might that argument go wrong? Well, to begin with, there will occasionally be

times and places—such as Japan in the 1990s—where the need to boost aggregate

demand is extremely large and lasts a very long time, longer than any conceivable lags in

fiscal policy. Models with hysteresis offer extreme examples of this possibility, but a

model with a maximum root of 0.98 or so will do almost as well. And remember, Stock

and Watson (1999) estimated the maximum root for real GDP in the United States to be

between 0.96 and 1.10.

Second, there are institutional structures that can make the inside lags in fiscal policy

quite short. The U.K.’s parliamentary system legislates fiscal changes very quickly, and

even the U.S. Congress has shown itself capable of moving within a few months—on

occasion! It may be that we generalized too glibly from the terribly long inside lags

witnessed in 1968 and 1974. Furthermore, if we really want to speed up fiscal policy

decisionmaking, there are a variety of institutional changes that could help.

Third, to the extent that we are worried that low MPCs out of temporary income tax

changes will undermine the effectiveness of fiscal policy, there are (non-income) tax

options that can induce intertemporal substitution by reducing current prices relative to

future prices—for both consumer goods and investment goods.

In summarizing the case against fiscal stabilization, Feldstein (2002) added one further item to this list:

the possibility that tax cuts or expenditure increases can depress demand by raising long-term interest rates.

But he did not suggest that tax increases or expenditure cuts can stimulate the economy by lowering

interest rates.

Fourth, but these more exotic options may not even be needed, for a fascinating body

of recent econometric evidence suggests that a sizable fraction of the U.S. population

(even weighted by income) is, or acts as if it is, subject to binding liquidity constraints.

Thus, even explicitly temporary changes in income taxes may pack significant spending

punch. Furthermore, with a little ingenuity we can target tax cuts on people who are

more likely to be living hand to mouth, such as poor people and the unemployed.

Fifth, there will occasionally be extraordinary circumstances—contemporary Japan is

again the outstanding example—where the zero bound on nominal interest rates makes

monetary policy a less powerful instrument of stabilization than it usually is. In such a

situation, monetary policy alone may be too weak to do the job, and a combined

monetary-fiscal effort—deficit spending or tax cuts financed by printing money—may be

needed. Indeed, fiscal policy might well be the senior member of such a partnership,

since a liquidity trap not only reduces the power of monetary policy but also increases the

power of fiscal policy (because there is little or no “crowding out” from higher interest

rates). Precisely this sort of two-pronged stabilization policy is what many economists

long urged on Japan.

Sixth, in certain rare “emergencies”—such as the U.S. in the aftermath of the 9/11

attacks—the monetary policy medicine may simply be too slow-acting to provide a

timely cure. The inside lags in monetary policy would probably be negligible in a clear

emergency.

But the outside lags remain quite long—a year or more for real GDP, and

two years or more for inflation—and there is not much the Federal Reserve can do about

it. With monetary policy lags as long as that, fiscal policy may be the only cyclical

medicine that can work in time—provided (and this is truly a big “if”) the inside lags can

For example, the Federal Reserve cut interest rates within days of the 9/11 attacks.

be kept short. In fact, in the end, I am inclined to conclude that the long inside lags (with

the concomitant politics) constitute the most important count in the indictment against

fiscal policy.

So my overall conclusion runs something like this. Under normal circumstances,

monetary policy is a far better candidate for the stabilization job than fiscal policy. It

should therefore take first chair. Nothing in this paper is intended to dispute this piece of

conventional wisdom. That said, however, there will be occasional abnormal

circumstances in which monetary policy can use a little help, or maybe a lot, in

stimulating the economy—such as when recessions are extremely long and/or extremely

deep, when nominal interest rates approach zero, or when significant weakness in

aggregate demand arises abruptly.

To be prepared for such contingencies, it makes

sense to keep one or more fiscal policy vehicles tuned up and parked in the garage, and

perhaps even to adopt institutional structures that make it easier to pull them out and take

them for a spin when needed.

As previously noted, I see little hope that fiscal policy can be used effectively in the contractionary

direction.

BIBLIOGRAPHY

Auerbach, Alan J. and Kevin Hassett. 1991. “Recent U.S. Investment Behavior and the

Tax Reform Act of 1986: A Disaggregate View.” Carnegie-Rochester Conference Series

on Public Policy 35, Autumn: 185-215.

Barro, Robert J. 1974. “Are Government Bonds Net Wealth?” Journal of Political

Economy 82(6), November/December: 1095-1117.

Barro, Robert J. 1979. “On the Determination of the Public Debt.” Journal of Political

Economy 87(5), October: 940-971.

Barro, Robert J. 1981. “Output Effects of Government Purchases.” Journal of Political

Economy 89(6), December: 1086-1121.

Barro, Robert J. 1986. “The Behavior of United States Deficits.” In The American

Business Cycle: Continuity and Change, edited by Robert J. Gordon. National Bureau of

Economic Research Studies in Business Cycles 25: 361-387. University of Chicago

Press.

Barsky, Robert B., N. Gregory Mankiw, and Stephen P. Zeldes. 1986. “Ricardian

Consumers with Keynesian Propensities.” American Economic Review 76(4), September:

676-691.

Bernanke, Ben S. 2000. “Comment on America’s Historical Experience with Low

Inflation.” Journal of Money, Credit, and Banking 32(4), November, part 2: 994-997.

Blanchard, Olivier J. 1984. “Current and Anticipated Deficits, Interest Rates and

Economic Activity.” European Economic Review 20(1), June: 7-27.

Blanchard Olivier, and Roberto Perotti. 2002. “An Empirical Characterization of the

Dynamic Effects of Changes in Government Spending and Taxes on Output.” Quarterly

Journal of Economics 117(4), November: 1329-1368.

Blinder, Alan S. 1981. “Temporary Income Taxes and Consumer Spending.” Journal of

Political Economy 89(1), February: 26-53.

Blinder, Alan S. 1997. “Is Government Too Political?” Foreign Affairs 76(6),

November/December: 115-126.

Blinder, Alan S. 2001. “The Economic Stimulus We Need.” The New York Times,

September 28.

Blinder, Alan S., and Angus Deaton. 1985. “The Time Series Consumption Function

Revisited.” Brookings Papers on Economic Activity 2: 465-511.

Blinder, Alan S., and Robert M. Solow. 1974. “Analytical Foundations of Fiscal Policy.”

In The Economics of Public Finance, edited by Alan S. Blinder and others. Washington:

Brookings Institution.

Blinder Alan S., and Janet L. Yellen. 2001. The Fabulous Decade: Macroeconomic

Lessons from the 1990s. New York: Century Foundation Press.

Business Council of Australia. 1999. “Avoiding Boom/Bust: Macroeconomic Reform in

a Globalised Economy.” New Directions Discussion Paper 2.

Campbell, John Y., and N. Gregory Mankiw. 1987. “Are Output Fluctuations

Transitory?” Quarterly Journal of Economics 102(4), November: 857-880.

Canzoneri, Matthew B., Robert E. Cumby, and Behzad T. Diba, 2002. “Should the

European Central Bank be Concerned about Fiscal Policy?” Rethinking Stabilization

Policy. Federal Reserve Bank of Kansas City. Pages 333-389.

Chirinko, Robert S. 1993. “Econometric Models and Empirical Findings for Business

Investment.” Salomon Brothers Center Series Financial Markets, Institutions &

Instruments Monograph. New York: Basil Blackwell.

Chirinko, Robert S. 1999. “Investment Tax Credits.” In Encyclopedia of Taxation and

Tax Policy, edited by in Joseph J. Cordes, Robert D. Ebel, and Jane G. Gravelle. New

York: Urban Institute Press (for the National Tax Association). Pages 211-215.

Congressional Budget Office. July 2003. “How CBO Analyzed the Macroeconomic

Effects of the President’s Budget.” Washington: U.S. Government Printing Office.

Deaton, Angus. 1992. Understanding Consumption. Oxford University Press.

Eckstein, Otto. 1978. The Great Recession, with a Postscript on Stagflation. Amsterdam:

North-Holland.

Eisner, Robert. 1969. “Fiscal and Monetary Policy Reconsidered.” American Economic

Review 59(5), December: 897-905.

Elmendorf, Douglas W., and David L. Reifschneider. 2002. “Short-Run Effects of Fiscal

Policy with Forward-Looking Financial Markets,” National Tax Journal 55, September:

357-386.

Fatás, Antonio, Jurgen von Hagen, Andrew Hughes Hallett, Anne Sibert, and Rolf R.

Strauch. 2003. Stability and Growth in Europe: Towards a Better Pact. Monitoring

European Integration 13. London: Centre for Economic Policy Research.

Feldstein, Martin S. 2001. “Japan Needs to Stimulate Spending.” The Wall Street

Journal, July 16.

Feldstein, Martin S. 2002. “Commentary: Is There a Role for Discretionary Fiscal

Policy?” Rethinking Stabilization Policy. Federal Reserve Bank of Kansas City. Pages

151-162.

Friedman, Milton. 1968. “The Role of Monetary Policy.” American Economic Review

58:1-17.

Gordon, Robert J., ed. 1986. The American Business Cycle: Continuity and Change.

University of Chicago Press.

Gordon, Robert J. 1980. “Postwar Macroeconomics: The Evolution of Events and Ideas.”

In The American Economy in Transition, edited by Martin S. Feldstein. National Bureau

of Economic Research Monograph. University of Chicago Press. Pages 101-162.

Hall, Robert E. 1978. “Stocastic Implications of the Life Cycle-Permanent Income

Hypothesis: Theory and Evidence.” Journal of Political Economy 86(6): 971-987.

Hall, Robert E., and Frederic S. Mishkin. 1982. “The Sensitivity of Consumption to

Transitory Income: Estimates from Panel Data on Households.” Econometrica 50(2),

March: 461-481.

Heller, Walter W. 1966. New Dimensions of Political Economy. Harvard University

Press.

Hsieh, Chang-Tai. 2003. “Do Consumers React to Anticipated Income Changes?

Evidence from the Alaska Permanent Fund.” American Economic Review 93(1): 397-405.

Jappelli, Tullio. 1990. “Who is Credit Constrained in the U.S. Economy?”, Quarterly

Journal of Economics 105(1): 219-234.

Johnson, David S., Jonathan A. Parker, and Nicholas S. Souleles. 2004. “The Response

of Consumer Spending to the Randomized Income Tax Rebates of 2001.” Princeton

University Working Paper (February).

Keynes, John M. 1936. The General Theory of Employment, Interest and Money. New

York: Harcourt, Brace.

Krueger, Alan B. 2001. “Temporary Stimulus Needed.” The New York Times, October

13: C3.

Krugman, Paul R. 1998. “It’s Baaack: Japan’s Slump and the Return of the Liquidity

Trap.” Brookings Papers on Economic Activity 2: 137-187.

Lerner, Abba P. 1943. “Functional Finance and the Federal Debt.” Social Research 10:

38-51.

Lindbeck, Assar, and Dennis J. Snower. 1988. The Insider-Outside Theory of

Employment and Unemployment. MIT Press.

Lucas, Robert E. 1976. “Econometric Policy Evaluation: A Critque.” Carnegie-Rochester

Conference Series on Public Policy 1: 19-46.

Modigliani, Franco, and Charles Steindel. 1977. “Is a Tax Rebate an Effective Tool for

Stabilization Policy?” Brookings Papers on Economic Activity 1: 175-203.

Musgrave, Richard A. 1959. The Theory of Public Finance. New York: McGraw-Hill.

O’Driscoll, Gerald P. Jr. 1977. “The Ricardian Nonequivalence Theorem.” Journal of

Political Economy 85(1), February: 207-210.

Okun, Arthur M. 1970. The Political Economy of Prosperity. New York: W.W. Norton.

The Brookings Institution.

Okun, Arthur M. 1971. “The Personal Tax Surcharge and Consumer Demand 1968-70.”

Brookings Papers on Economic Activity 1: 167-211.

Parker, Jonathan A. 1999. “The Reaction of Household Consumption to Predictable

Changes in Social Security Taxes.” American Economic Review 89(4), September: 959-

973.

Patinkin, Don. 1956. Money, Interest, and Prices: An Integration of Monetary and Value

Theory. New York: Harper & Row.

Phelps, Edmund S. 1968. “Money-Wage Dynamics and Labor Market Equilibrium.”

Journal of Political Economy 76: 678-711.

Samuelson, Paul A. 1948. Economics: An Introductory Analysis. New York: McGraw-

Hill.

Seidman, Laurence S. 2003. Automatic Fiscal Policies to Combat Recessions. Armonk,

New York: M.E. Sharpe.

Shapiro, Matthew D., and Joel Slemrod. 1995. “Consumer Response to the Timing of

Income: Evidence from the Change in Tax Withholding.” American Economic Review

85(1), March: 274-283.

Shapiro, Matthew D., and Joel Slemrod. 2003. “Consumer Response to Tax Rebates.”

American Economic Review 93(1), March: 381-396.

Solow, Robert M. 1966. "The Case Against the Case Against the Guideposts." In

Guidelines, Informal Controls and the Market Place, edited by George Shultz and Robert

Aliber. Chicago: University of Chicago Press. Pages 41-54.

Solow, Robert M. 2002. “Is Fiscal Policy Possible? Is It Desirable?” Presidential address

to the XIII World Congress of the International Economic Association, Lisbon,

September.

Souleles, Nicholas S. 1999. “The Response of Household Consumption to Income Tax

Refunds.” American Economic Review 89(4), September: 947-958.

Souleles, Nicholas S. 2002. “Consumer Response to the Reagan Tax Cuts.” Journal of

Public Economics 85(1), July: 99-120.

Stein, Herbert. 1969. The Fiscal Revolution in America. University of Chicago Press.

Stock, James H., and Mark Watson. 1999. “Business Cycle Fluctuations in U.S.

Macroeconomic Time Series.” In Handbook of Macroeconomics, vol. 1A, edited by John

B. Taylor and Michael Woodford. Amsterdam: North-Holland. Pages 3-64, Chapter 1.

Sumner, M.T. 1979. “A Skeptical Note on the Efficacy of Temporary Sales Tax

Reductions.” Canadian Public Policy 5(1), Winter: 97-101.

Turnovsky, Stephen J., and Marcus H. Miller. 1984. “The Effects of Government

Expenditure on the Term Structure of Interest Rates.” Journal of Money, Credit and

Banking 16(1), February: 16-33.

U.S. Treasury. 1979. “The Annual Report of the Secretary of the Treasury on the State of

the Finances.” Washington: U.S. Government Printing Office.

von Hagen, Jurgen, and Ian J. Harden. 1995. “Budget Processes and Commitment to

Fiscal Discipline.” European Economic Review 39(3-4), April: 771-779.

Wyplosz, Charles. 2002. “Fiscal Discipline in EMU: Rules or Institutions?” Working

Paper prepared for the Group of Economic Analysis of the European Commission

(April).

Zeldes, Stephen P. 1989. “Consumption and Liquidity Constraints: An Empirical

Investigation.” Journal of Political Economy 97(2), April: 305-346.