Skip to main content

Principles for Fiscal Stimulus Economic Policy in a Weakening Economy

Principles for Fiscal Stimulus

The weak December jobs report that the Labor Department issued January 4 is the latest evidence the economy has weakened significantly and is now growing quite slowly. A number of economists continue to believe this slow growth will continue into 2008, but that the economy will avoid slipping into a recession.[1]


Some prominent economists are more pessimistic. For example, former Treasury Secretary Lawrence Summers and Martin Feldstein, chairman of President Reagan’s Council of Economic Advisers, place the odds of a recession in 2008 at 50-50 or higher and have urged Congress to begin taking steps now to be prepared with a fiscal stimulus package.[2]

Congress should weigh its options very carefully and not be stampeded into moving stimulus legislation as an antidote to temporary economic weakness. Substantial uncertainty remains about whether economic conditions will deteriorate sufficiently to warrant additional fiscal stimulus, on top of the stimulus that monetary policy can provide. Decisions by Congress and the President regarding whether to institute fiscal stimulus measures need to be based on emerging data on the economy, the actions the Federal Reserve takes, and the decidedly mixed historical record of fiscal stimulus actions in the past. It is particularly important that concern over a possible recession not be used to justify fiscally irresponsible measures that would exacerbate the nation’s already serious long-term fiscal problems without providing cost-effective short-run stimulus.

On the other hand, if Congress waits to begin assembling a stimulus package until data are available demonstrating that the economy is in a full-blown recession, then fiscal stimulus may come too late to do much good.

Congress can respond to these competing concerns by beginning work now to fashion a package of effective stimulus measures, while making its actual implementation contingent upon further evidence of economic weakness. In this regard, Martin Feldstein’s proposal to enact a contingent stimulus package quickly — and to use a decline in employment over a three-month period as a “trigger” for putting it into effect — merits consideration. Under such an approach, Congress and the President would enact a stimulus package soon, and the package would take effect if private payroll employment in the U.S. economy in any month fell below the level it had attained three months earlier.

More important than the choice of a trigger for a stimulus package, however, is the design of the package itself. Various measures billed as “stimulus” can entail substantial cost while doing little to stimulate the economy if they are designed poorly, in response to political pressures or ideological fixations. Analysts recognize, as Lawrence Summers has said, that to be effective, stimulus measures must meet three basic tests: they must be timely, targeted, and temporary.

  • Timely measures are those that, once triggered, stimulate new spending quickly so that businesses do not have to cut back on production or lay off workers due to weak demand. Increased funding for government programs that spend-out slowly, or for tax cuts that will be saved in large part rather than spent quickly, are poor candidates for stimulus; they have little impact when the need is greatest and only kick in when the economy is reviving anyway.
  • Targeted measures are those aimed at individuals and entities that will spend quickly the bulk of any new resources they receive. Tax cuts that mainly benefit high-income individuals are poorly targeted to provide stimulus, because those individuals are more likely to save a large share of any increase in disposable income they receive than are people of more modest means. Government-funded construction projects that take many months or even several years to get underway are poorly targeted as well. In contrast, tax cuts and increases in government spending aimed at low- and moderate-income consumers and unemployed workers — such as tax cuts that provide a flat refund to all tax filers, additional weeks of unemployment benefits to workers who have been unable to find a new job, and increases in food stamp benefits — are far more effective as stimulus. Fiscal relief for state governments is another well-targeted form of stimulus. Faced with declining tax receipts and looming deficits when the economy slows, states respond by raising taxes and cutting back on spending, especially for health care, education, and aid to local governments. States have little alternative, since they are required to balance their budgets in bad economic years as well as good ones. But such actions further reduce demand in the economy and deepen the recession. Federal fiscal relief that enables states to minimize such spending cuts and tax increases thus helps prop up the economy in a time of weakness.
  • Temporary measures are those that expire once the economy improves. The country should not be stuck with permanent, deficit-increasing tax cuts or spending increases because of a temporary economic downturn; it is essential that all stimulus measures terminate when the economy strengthens. It also is desirable that short-term increases in the budget deficit caused by a stimulus package be offset by deficit-reduction measures that take effect in subsequent years. Otherwise, stimulus measures will cause some permanent increases in interest payments on the debt that will worsen long-term deficits, although the effects should be small if the stimulus measures are truly temporary.

Some policymakers appear to assume that tax cuts are inherently stimulative, while spending increases are inherently less desirable as economic stimulus. Such assumptions do not withstand scrutiny. Both spending measures and tax cuts can be effective — or ineffective — as stimulus, depending on their nature and design. Nobel laureate Joseph Stiglitz and now-CBO director Peter Orszag wrote in late 2001, “Basic economic analysis indicates that increased government expenditures can indeed be stimulative, and, in fact, are often more effective as stimulus measures than tax cuts.”[3] Similarly, two senior Federal Reserve economists found in 2002 that increases in government spending tend to have a greater stimulative effect than tax cuts that have the same cost, because more of the increase in government spending will translate quickly into an increase in total spending in the economy, while a substantial part of a tax cut will generally be saved.[4]

This analysis assesses the risks and benefits of acting on fiscal stimulus now and presents principles and recommendations for the design of an effective stimulus package.

The Case for Fiscal Stimulus Is Not a Slam-Dunk

Economic stimulus is warranted when the major problem facing the economy is “economic slack” — weak aggregate demand relative to the economy’s capacity to produce goods and services — that results in excess unemployment and unused productive capacity. The goods and services not produced and the income lost during periods of economic slack are lost forever. The primary objective of stimulus policy is to minimize such losses.

When there is little economic slack, however, stimulus is unnecessary and can be counterproductive. Instead of producing sustainable increases in the production of goods and services, such stimulus can increase inflationary pressures by raising aggregate demand above the economy’s capacity to meet it.

While the most recent economic data are troubling, they do not yet provide evidence that there is so much economic slack that actions by the Federal Reserve could not provide sufficient stimulus to avert a recession. To be sure, the recovery from the 2001 recession has been weak by historical standards, and the gains from economic growth since 2001 have been skewed toward corporate profits and the incomes of highly compensated individuals.[5] But the most current data now available on real economic growth — 3.8 percent at an annual rate in the second quarter of 2007 and 4.9 percent in the third quarter — are well above the 3¼ -3½ percent annual rate generally believed to be sustainable over the long term. In addition, while the jump in the unemployment rate to 5.0 percent in December 2007 was substantial, the unemployment rate is still only a little above the 4.8 percent average rate that the Congressional Budget Office (CBO) projects for the 2009-2017 period.

Of course, these data are backward looking. The reason that stimulus is now on the table for discussion is that forecasters generally believe the economy has already slowed substantially from its third-quarter growth rate. The latest employment data show that job growth ground to a halt in December (18,000 jobs added overall, but a decline of 13,000 jobs in the private sector). And near-term projections are for the economy to grow at a pace well below its long-run growth rate, in the face of ongoing adjustments in the housing market, declining consumer confidence, and high oil prices. The most serious concern is that financial markets generally could seize up due to contagion from the subprime mortgage crisis. Were that to happen, businesses and households would have trouble financing their spending, and the economy could fall into recession that would require a stronger policy response than the Fed alone could provide.

To stimulate spending, the Federal Reserve has begun cutting interest rates and easing money and credit conditions. The Fed has substantial room to cut interest rates further; the monetary authority’s target for the federal funds rate is now 4.25 percent, well above the 1 percent target rate the Fed maintained from mid-2003 to mid-2004 to support the recovery from the 2001 recession. At its most recent meeting on December 11, the Fed’s attitude was cautious. Citing uncertainty about both economic growth and inflation, the Fed said that going forward, it “will act as needed to foster price stability and sustainable economic growth.” In the wake of the December employment report, however, the Fed is almost surely leaning more toward lowering interest rates, and many forecasters are looking for a half-point cut in the federal funds target when the Fed meets again at the end of this month.

Under normal circumstances, the job of taking explicit action to stabilize fluctuations in economic activity is left to monetary policy since, as CBO director Peter Orszag recently told Congress, fiscal stimulus packages enacted by Congress and the President often “prove to be poorly timed, or designed in a relatively ineffective way” to do much good in a recession.[6] The Fed, in contrast, can make changes in monetary policy immediately upon deciding they are needed.

Moreover, fiscal policy already acts automatically to help stabilize the economy — not through one-time tax cuts or spending increases contained in a stimulus package, but through what are known as the “automatic stabilizers” — the reductions in tax payments and increases in government expenditures for programs like unemployment insurance that take place automatically when a recession sets in. In particular, a decline in economic activity automatically produces a proportionally greater decline in tax receipts, cushioning the decline in people’s disposable incomes.

How an Employment Trigger Would Work

At the beginning of each month, the Bureau of Labor Statistics reports on the level of employment for the previous month, including revisions to data for earlier months. The change in private payroll employment is probably the best and most timely official data that could be used to trigger a stimulus package.

Martin Feldstein has recommended such a trigger, under which stimulus measures would take effect if there is a decline in payroll employment over a three-month period as a whole. For example, if the employment report issued at the start of April 2008 showed that private payroll employment was lower in March 2008 than in December 2007, that would strongly suggest the economy was entering a recession and the enacted stimulus measures would automatically take effect. This would be a prudent course in the present situation, where uncertainty remains about the economy’s near-term prospects.

No measure is a perfect real-time predictor of recession. However, a three-month private employment decline is about as good a measure as one can get. The measure would have triggered relief in March 2001 (the exact month the 2001 recession began), May 1990 (two months before the 1990-91 recession began), and October 1981 (three months into the severe 1981-82 recession). When the economy declines, data showing a three-month decline in payroll employment become available many months before data showing that growth in the Gross Domestic Product has been negative for two consecutive calendar quarters, which is a common informal indicator that a recession is underway.

Accordingly, a stimulus package is warranted only if the decline in economic activity is, or is expected to be, sufficiently large that monetary policy alone (supported by the automatic stabilizers) will not provide sufficient offsetting stimulus. At this point, the Federal Reserve apparently is not convinced that the evidence that the economy is weakening is strong enough to conclude that a recession is imminent, although the case for an easier monetary policy is stronger than it was before the December jobs report came out.

Congress’s current options thus are to: a) largely eschew a fiscal stimulus package; b) enact a stimulus package to go into effect immediately, on the assumption that the Fed will not be able to act quickly or strongly enough; or c) enact a stimulus package quickly, but with a trigger based on economic conditions so that the package would go into effect rapidly if the evidence shows it is needed. In light of the considerable uncertainty about where the economy is headed and how well it will respond to Fed actions, the third course — with an employment-related trigger such as that suggested by Martin Feldstein (see the box above) — represents the soundest course (assuming, of course, that the package adheres to the principles for sound stimulus policy laid out below and is fiscally sound over the longer term).

What Principles Should Guide the Design of a Stimulus Package?

As stated above, a truly effective stimulus policy should be timely, targeted, and temporary, as well as fiscally responsible over the long term. In the fall of 2001, the chairmen and ranking members of the House and Senate Budget Committees, a bipartisan group that included the current OMB director, Jim Nussle, agreed to such principles in a joint statement that remains relevant today.[7]

Some Stimulus Measures Have Much More “Bang for the Buck” Than Others

A good way to assess a stimulus proposal is through its fiscal “bang for the buck” — how much immediate spending boost it would deliver for each dollar it costs. During the debate in 2002-2003 over a stimulus package, economist Mark Zandi of evaluated various stimulus options in terms of their effect on the economy in the first year after enactment.

While economic conditions have changed since then, Zandi’s rankings continue to provide a useful assessment of the relative value of various measures in generating economic stimulus. They indicate that temporarily strengthening unemployment insurance is among the most effective forms of stimulus, because most of the additional unemployment benefits would be quickly spent, while a dividend or capital gains tax cut would be among the most ineffective measures because much of it would be saved rather than consumed quickly.

Demand generated per $1 of cost
High “Bang for the Buck”
Extend unemployment benefits $1.73
Provide state fiscal relief $1.24
Enact a one-time uniform tax rebate $1.19
Increase Child Tax Credit $1.04
Lower “Bang for the Buck”
Adjust Alternative Minimum Tax exemption levels $0.67
Reduce marginal tax rates $0.59
Increase tax breaks for small business investment $0.24
Cut taxes on dividends and capital gains $0.09
Reduce estate tax $0.00

Source: Mark M. Zandi, “Assessing President Bush’s Fiscal Policies,”, July 2004.; “The Economic Impact of the Bush and Congressional Democratic Economic Stimulus Plans,”, February 2003.

They agreed that stimulus should be rapid, temporary (lasting a year at most), well targeted, and appropriately sized — and that its cost should be fully offset in subsequent years. (The statement noted that “an economic stimulus package should be based on the recognition that long-term fiscal discipline is essential to sustained economic growth.”) Today, many economists across the political spectrum concur that stimulus should adhere to these core principles.

A brief discussion of these core principles follows.

Stimulus Needs To Be Timely and Well-Targeted

To boost demand for goods and services, fiscal stimulus should be targeted where it is most likely to be spent immediately and thereby to boost demand. This suggests targeting on low- and moderate-income individuals and unemployed workers who need a replacement for lost income. People whose income is disrupted in a recession and who lack the savings to tide them over and maintain their normal consumption, and people whose incomes are so low to start with that they have difficultly making ends meet, are the people most likely to spend quickly any added income they receive. Assisting them thus provides strong stimulus.

The need for stimulus to be targeted where it will be spent quickly also suggests targeting on state governments, which will need federal fiscal assistance to avert substantial tax increases and spending cuts, since they are required to balance their budgets even during recessions. In contrast, stimulus directed toward long-term spending such as new construction projects, tax cuts for high-income households, or cuts in business tax rates would be neither timely nor well-targeted.

Low-Income Households

Because low- and moderate-income individuals often have difficulty paying ongoing household costs, they are likely to spend quickly any additional income they receive. Wealthier individuals, in contrast, are more inclined to save a substantial share of additional income they receive, because they do not need it to meet their immediate needs.

This distinction is crucial. CBO has noted that “the efficacy of fiscal stimulus depends critically on households’ tendency to spend the income placed in their hands.” The larger the share of income that a group of households spends, “the more powerful is the ultimate boost to demand” from stimulus measures that increase those households’ incomes.[8]

Relief for low-income households usually is best administered through spending programs rather than the tax code. Because many low-income families do not (and are not required to) file tax returns, stimulus delivered through the income tax code would miss them entirely.

It also is more efficient to provide economic stimulus through existing channels and without creating new bureaucracy. For example, CBO director Peter Orszag recently stated that “certain types of spending, especially [those that] involve transfer payments, for example, food stamps, … do boost demand for goods and services pretty quickly and pretty effectively in the short run.”[9] (An increase in food stamp benefits can begin to reach low-income households and to be injected into the economy within about 60 days of enactment, with virtually no increase in bureaucracy or administrative costs.) For such reasons, Martin Feldstein recently joined those calling for a temporary food stamp increase, noting that it would be stimulative because it would provide resources to people with a high propensity to consume.

Unemployed Workers

The unemployment insurance (UI) program is a joint federal-state program that provides partial, temporary income replacement to workers who have a demonstrated attachment to the labor force and who lose their jobs due to a layoff or other economic reasons or must otherwise leave their jobs through no fault of their own.

UI serves as an important automatic stabilizer. When the economy weakens, the unemployment rate rises and total UI payments increase. When the economy improves enough to lower the unemployment rate, UI payments decline.[10]

Temporary expansions in unemployment insurance benefits during a recession are particularly effective stimulus because they are both well-targeted and temporary. They go to workers who are involuntarily unemployed and whose income has fallen below their normal level of spending, a group that tends to be concentrated in the areas and industries most affected by the slowdown. As Joseph Stiglitz wrote in late 2001 with reference to proposals to temporarily expand UI benefits at that time, “give money to people who have lost their jobs in this recession, and it would be quickly spent.”[11]

Similarly, Peter Orszag observed in November 2001:

Temporary expansions in unemployment insurance help to break the downward economic spiral created by job layoffs by providing benefits to families of unemployed workers. Because the spending needs of these families typically exceed their income following the loss of a job, the families are likely to spend a high percentage of any additional income they receive during their periods of unemployment….

Unemployment insurance thereby promotes additional spending by households with unemployed workers, boosting demand for products and protecting the jobs of workers in the firms that produce those goods and services. Temporary expansions in unemployment insurance consequently are a “win-win” proposition: They are quite effective at helping more people to keep their jobs during an economic downturn, while also helping those who are unfortunate enough to lose their jobs.[12]

In a recent House Budget Committee hearing, Orszag reiterated the value of additional UI benefits as stimulus, commenting that “research has shown that the unemployment insurance system is among the most effective dollar-for-dollar economic stabilizers that we have in terms of counterbalancing periods of economic weakness.”[13]

Unemployed workers who qualify for UI benefits typically are eligible for 26 weeks of benefits. The UI system also includes an “extended benefits” program that was originally intended to provide additional weeks of benefits during downturns to workers in high-unemployment states who are unable to find a new job before their benefits run out. The trigger for the extended benefits works poorly, however, and in every recession of recent decades, the federal government has created a separate, temporary program to provide additional weeks of unemployment benefits to workers who have exhausted their regular UI benefits and continue looking for work. These federal extended benefits, along with some other needed UI reforms (see the discussion on page 13), are an important component of any fiscal stimulus package.

State Governments

Because states are required to balance their budgets each year, the drop in revenues that results from an economic slowdown can cause serious problems. The two major options for states to close their budget gaps — raising taxes and cutting spending — put further stress on a weakening economy, decreasing either the amount of money that people have to spend or the amount of money the state spends. The provision of temporary fiscal relief to state governments can enable states to minimize tax increases and budget cuts, which would be economically damaging during a recession.[14]

As of mid-December 2007, 19 states already were reporting that they face budget shortfalls for fiscal year 2009 (which begins for most states on July 1, 2008).[15] Thirteen of these states have quantified their projected shortfalls, which total at least $23 billion for 2009 for these states alone. Moreover, the number of states facing shortfalls is expected to rise markedly in coming weeks as more states complete budget reviews and governors unveil their 2009 budgets. And as the executive director of the National Conference of State Legislatures recently warned, state finances will deteriorate further if a recession occurs.[16]

Adding to these concerns, the bursting of the housing bubble threatens to stress state and local budgets well beyond what is typical in an economic slowdown. States are now forecasting lower state sales tax revenues due to significantly lower tax collections from the sales of furniture, appliances, construction materials, and other housing-related expenditures. Of even greater concern, the decline in housing values is beginning to translate into substantially reduced property tax revenues for many local governments.

For all of these reasons, federal relief to the states would constitute effective stimulus. The stimulus package enacted in 2003 included $20 billion in state fiscal relief, with half of it taking the form of a temporary, across-the-board increase in federal Medicaid matching rates (i.e., in the share of Medicaid costs that the federal government bears) and the other half taking the form of a block grant. If there is a downturn, states will need federal relief to avert, or mitigate the severity of, harsh service cuts or substantial tax increases at a time when such actions would weaken the economy further.[17] General aid to state governments would also reduce the likelihood that states would cut back on infrastructure spending included in their operating budgets, such as school repairs.[18]


During an economic downturn, aggregate demand lags, and existing capacity is not fully utilized. Companies that face decreased demand have little incentive to expand their capacity, regardless of available tax incentives. While some types of temporary tax incentives might encourage some firms to make investments quickly — that is, before the incentives expire — firms typically hesitate to increase investment when they lack confidence in their ability to sell their products.

A temporary investment incentive targeted to new investment (as distinguished from investments that have already been made) may provide stimulus in situations where temporary weakness in the economy has caused firms to postpone planned investments, if the incentive succeeds in inducing firms to accelerate their investment plans. However, the stimulative effects of such incentives are likely to be more modest than measures that put the same amount of money in the pockets of households who will spend it, because a substantial fraction of the investment benefiting from tax incentives would likely have been made anyway and would therefore not represent additional demand.[19] A study by two Federal Reserve economists of the effects of the “bonus depreciation” investment incentives enacted in 2002 and expanded in 2003 to provide stimulus during the last recession found that based on the available evidence, this measure had at best “only a very limited impact” on investment spending.[20].

Moreover, general reductions in business taxes such as a reduction in corporate income tax rates are largely ineffective as stimulus. Businesses base their hiring and firing decisions (and their decisions on purchasing raw materials) primarily on expectations about their ability to sell their goods and services — i.e., on customer demand — rather than on how much cash they have on hand.

A recent Goldman-Sachs analysis made this point, noting that “companies don’t spend money just because it’s there to spend. To justify outlays for new projects, the expected returns have to exceed the costs, and that usually requires growth in demand strong enough to put pressure on existing resources.”[21]

It is critical that any investment incentives included in a stimulus package be strictly temporary if the incentives are to provide meaningful stimulus. If firms can postpone investment decisions until consumer demand strengthens and still get the new investment tax breaks, they will be inclined to defer those investment decisions, thereby rendering the tax incentives ineffective as stimulus.

Stimulus Should Not Compromise Long-Term Economic Growth

Economic stimulus should not significantly worsen the long-term budget picture. The purpose of stimulus measures is to speed the economy’s recovery by boosting demand temporarily, and it is essential that stimulus measures expire after the slowdown is mitigated. As CBO warned when it evaluated tax-based stimulus options in 2002, if tax cuts enacted to provide short-term stimulus are maintained permanently, they can detract from long-term economic growth by increasing deficits.[22] Short-term economic stimulus is most effective when it remains just that — short-term policy.

Because stimulus measures are intended to boost total spending in the economy in the short run, they increase the budget deficit in the year in which they take effect. And just as state actions to cut spending or raise taxes would stress a faltering economy, so too would it be counterproductive to require that federal stimulus measures be offset over the same time period by deficit-reducing measures. Rigid adherence to a deficit-neutrality requirement on a year-by-year basis would be unwise.

It would be desirable, however, for stimulus measures that increase the budget deficit in the short run to be offset by deficit-reducing measures that take effect later on, when the economy has recovered. Doing so would be consistent with the PAYGO rules the House and Senate established last year. Those rules require not that tax cuts or entitlement increases be offset every year, but instead that such legislation not increase the cumulative deficit over the next six or 11 years. Such an approach also would be consistent with the principles that the bipartisan leadership of the House and Senate Budget Committees issued in 2001, when they said that “outyear offsets should make up over time for the cost of near-term economic stimulus.”[23] Actions to offset the costs of a stimulus package after the economy has recovered would not reduce the short-term stimulative effects but would improve the long-term budget picture.[24]

The most important step to ensure that stimulus legislation does not compromise long-term growth, however, is to make sure that all stimulus measures are strictly temporary and terminate when the economy recovers. Measures that stimulate the economy in the short run usually harm long-term growth if maintained for too long. Thus, if offsetting the costs of stimulus measures after the economy rebounds is desirable, making all stimulus measures strictly temporary is essential. (Consistent with this principle, Martin Feldstein has noted that the contingent tax cut he recommends should not be “a net permanent tax cut, but something counter-cyclical.”)

It may be noted that in the first half of the 1990s, when adherence to fiscal discipline was at its strongest, PAYGO rules were scrupulously adhered to — except for one occasion during the recession of that period, when Congress chose not to offset the provision of additional weeks of unemployment benefits as a stimulus measure. Significant long-term fiscal damage did not result, because the extra weeks of benefits expired as scheduled and because Congress faithfully adhered to PAYGO for all other entitlement and tax legislation.

Evaluating Specific Stimulus Proposals

The following measures rank highly according to the criteria for good stimulus:

  • Strengthened unemployment insurance. As discussed, temporary increases in unemployment insurance benefits are a particularly effective form of economic stimulus. The benefits go to workers who have lost their jobs and suffered a temporary decline in their income relative to their normal expenditures. As a result, the added income is likely to be spent quickly. Analysis by economists Alan Auerbach and Daniel Feenberg concluded that per dollar of cost, the automatic stabilizer effects of the increase in unemployment insurance payments that occurs in a downturn are at least eight times as large as the effects of the reduction in tax collections that occurs.[25] A fiscal stimulus package accordingly should include a temporary measure to provide additional weeks of federally funded UI benefits for workers who exhaust their regular UI benefits before they can find work. Additional weeks of UI benefits have been provided in every recession in recent decades, and the case for them will be especially strong if a new recession sets in. The percentage of unemployed workers who have remained without a job for more than 26 weeks (the normal duration for regular unemployment benefits) and continue to search for work — a group sometimes referred to as the “long-term unemployed” — has remained stubbornly high and is considerably higher now than it was at the start of the last recession (17.5 percent in December 2007, compared with 11.1 percent in March 2001). In addition, UI benefit levels are quite low in many states, so a temporary increase in the weekly benefit amounts also would warrant consideration. In addition, Congress should take long-overdue action to address weaknesses in the UI program itself. UI coverage has eroded in recent decades, and only 37 percent of unemployed workers now receive unemployment benefits. The design of the UI program dates in substantial part from the 1930s; many low-income and part-time workers who are laid off do not qualify for UI benefits because of outmoded rules that do not reflect today’s workplace. (This is a particular problem for low-income women who lose their jobs.) In the mid-1990s, the congressionally chartered, bipartisan Advisory Commission on Unemployment Insurance recommended measures to remedy this problem, but Congress never acted on them. In October 2007, the House finally passed UI reform provisions[26] that reflect several key commission recommendations and would provide states with incentives to modernize their UI systems so more female, low-income, and part-time workers who lose their jobs through no fault of their own can qualify for benefits. (The measure is fully paid for through a renewal of the federal unemployment insurance surtax, a step that President Bush called for in his last budget.) If the Senate were to act expeditiously on these provisions or similar legislation in January, these reforms would be in place to strengthen UI’s automatic stabilizer role in the next recession, whenever that occurs. But if rapid Senate action on that legislation is not forthcoming, this problem could be addressed on a temporary basis as part of a stimulus package.
  • State fiscal relief. As noted, temporary federal fiscal relief reduces the need for states to enact tax increases and spending cuts in a recession. A temporary increase in the federal Medicaid matching rate (to help states avert cutting Medicaid coverage) is one effective component of such relief, along with the provision of general fiscal relief to states. In addition, in light of the growing pressures on local governments from declining property tax revenues, fiscal relief should be designed to encourage states to temporarily increase assistance to local governments, rather than to cut back such aid, as states often do in economic downturns to help balance their budgets. In short, as in 2003, temporary fiscal relief to states could include a combination of general aid and Medicaid-specific relief.
  • Uniform tax refunds. Tax cuts can provide effective economic stimulus, if — but only if — taxpayers spend quickly the resulting increase in their disposable incomes. This argues against proposals for across-the-board reductions in income tax rates or other tax breaks that would provide the largest benefits to higher-income taxpayers, who are likely to save rather than spend a substantial portion of any tax cut they receive. (At a recent Brookings Institution forum on stimulus, Martin Feldstein concurred that a temporary tax rate cut would not be effective as stimulus.) This also argues against tax cuts in capital gains or dividends, which would go even more disproportionately to those high up on the income scale. In contrast, a tax rebate of a uniform amount for all tax filers would put money into people’s hands quickly and direct a large proportion of the total to people likely to spend it. The rebate could go to all taxpayers who filed a federal income tax return the previous year, or to everyone who worked and paid FICA taxes (whether or not they filed an income tax return). The amount, presumably several hundred dollars, would be the same for all taxpayers and would not be tied to how much tax they paid.
  • Temporary increase in food stamp payments. Many low-income consumers are not tax filers and do not receive unemployment insurance. Hence, they would not qualify for a tax rebate or for extended UI benefits. Yet they experience particular hardship when the job market is weak and when states cut back on services. A temporary increase in food stamp benefit levels would provide helpful support to poor households and would almost certainly be spent very quickly. Dollar-for-dollar, this is one of the most effective forms of stimulus available.

At the other end of the spectrum, permanent cuts in marginal tax rates, such as making permanent the tax cuts enacted in 2001 or 2003, would score very low on the criteria for good fiscal stimulus. Not only would most of the benefits go to higher-income individuals, but extending tax cuts into 2011 and beyond would do little or nothing to boost consumer spending in 2008. Indeed, permanent tax cuts whose costs are not offset would do more to harm the economy over the long term, by increasing deficits, than to stimulate it in the short term. (Such tax cuts might even harm the economy in the short term as well, if the prospect of substantially increased long-term deficits caused long-term interest rates to be higher than they otherwise would be.) Finally, although increased investment in infrastructure has strong merit, new infrastructure projects (as distinguished from maintenance of projects already underway and possible acceleration of some planned repair projects that are ready to go and require little lead-time) do not rate well as stimulus, because infrastructure projects typically take many months to get off the ground.

Measures to Address Housing Market Problems Could be Useful But Should Be Assessed by Criteria Different From Those Used for Stimulus Measures

This paper has not discussed policies needed to address the fallout from problems in the housing market and the subprime mortgage crisis. That is not because these are not important policy concerns or are not tied to the immediate risks that the economy faces. They are. But most of the problems that have emerged because of the housing slump and the subprime mortgage crisis need to be addressed irrespective of whether the economy slides into recession. Furthermore, the criteria for evaluating those policies are different from the criteria that should be used when the primary objective of policies is to provide economic stimulus.

To be sure, to the extent that policies aimed at limiting the damage from mortgage foreclosures would support spending that would not otherwise take place, such policies can contribute to economic stimulus. But such measures should be considered on the basis of their suitability for addressing problems in the housing and mortgage markets, rather than as policies whose primary goal is to provide immediate stimulus to the economy.*

* For a thoughtful analysis of principles for addressing foreclosures, see Douglas W. Elmendorf, “What Should Be Done to Help Households Facing Foreclosure?”, The Brookings Institution, November 2007,


There is widespread consensus that the economy has weakened entering 2008, but considerable uncertainty about whether it is entering (or already has entered) a recession. Based on the available economic data, and given the considerable scope that remains for monetary stimulus by the Federal Reserve, it is not yet clear that fiscal stimulus measures are needed.

Because of the risk that economic conditions could worsen, however, Congress could act to get ready a stimulus package that could be put into effect quickly should the need arise. One way to do this would be to enact a contingent stimulus package that would take effect if a trigger — such as a decline in private employment over a three-month period — is pulled. Good candidates for effective stimulus measures include temporary increases in unemployment insurance, food stamps, and aid to state governments, as well as uniform tax rebates. In contrast, cutting income tax rates or taxes on capital gains or dividends, making the recent tax cuts permanent after 2010, and most infrastructure projects, score poorly as stimulus measures.

End Notes

[1] At a December 5 House Budget Committee hearing, Congressional Budget Office director Peter Orszag said, “Most analysts currently believe the economy will avoid a recession but will grow relatively slowly for several quarters.” “The Current Economic Situation,” Statement of Peter Orszag, Director, Congressional Budget Office, before the Committee on the Budget, U.S. House of Representatives, December 5, 2007, p. 1.

[2] See Lawrence Summers, “Why America Must Have a Fiscal Stimulus,” Financial Times, January 6, 2008,, and Martin Feldstein, Testimony before the Committee on the Budget, U.S. House of Representatives, December 5, 2007.

[3] Peter Orszag and Joseph Stiglitz, “Tax Cuts Are Not Automatically the Best Stimulus: A Response to Glenn Hubbard,” Center on Budget and Policy Priorities, November 27, 2001,

[4] Douglas W. Elmendorf and David L. Reifschneider, “Short-Run Effects of Fiscal Policy with Forward-Looking Financial Markets,” National Tax Journal, Vol. 55, No. 3, 2002.

[5] Aviva Aron-Dine, Chad Stone, and Richard Kogan, “How Robust Is the Current Economic Expansion?,” Center on Budget and Policy Priorities, revised October 16, 2007,; Aviva Aron-Dine and Isaac Shapiro, “Share of National Income Going to Wages and Salaries at Record Low in 2006: Share of National Income Going to Corporate Profits at Record High,” Center on Budget and Policy Priorities, revised March 29, 2007,

[6] “State of the U.S. Economy and Implications for the Federal Budget,” Hearing before the House Budget Committee, House of Representatives, December 5, 2007.

[7] “Principles for Economic Stimulus,” Kent Conrad, Pete Domenici, Jim Nussle, and John Spratt, October 4, 2001.

[8] Congressional Budget Office, “Economic Stimulus: Evaluating Proposed Changes in Tax Policy,” January 2002,

[9] “State of the U.S. Economy and Implications for the Federal Budget,” Hearing before the House Budget Committee, House of Representatives, December 5, 2007.

[10] Alan J. Auerbach and Daniel Feenberg,The Significance of Federal Taxes as Automatic Stabilizers," Journal of Economic Perspectives, Vol. 14, No. 3, 2000.

[11] Joseph Stiglitz, “A Boost That Goes Nowhere,” The Washington Post, November 11, 2001, page B01.

[12] Peter Orszag, “Strengthening Unemployment Benefits Would Be Much More Effective In Saving Jobs Than Most Corporate Tax Cuts,” Center on Budget and Policy Priorities, November 14, 2001,

[13] “State of the U.S. Economy and Implications for the Federal Budget,” Hearing before the House Budget Committee, House of Representatives, December 5, 2007.

[14] Some may ask whether, if the federal government provides aid to states in a recession, this will create a “moral hazard” problem in which states overspend, cut taxes too much, and/or fail to build up “rainy day” funds during periods of economic growth because they expect the federal government to bail them out when an economic downturn comes. The evidence suggests that the federal fiscal relief provided in the past did not have this effect.

The federal government provided such relief in the last downturn, and states have not overspent or slashed taxes since then in the expectation that they would be bailed out during future downturns. On average, state expenditures as a share of the economy are lower now than they were in state fiscal year 2001, while state taxes as a share of the economy are at about the same level. In addition, states built up substantial “rainy day” reserve funds to draw upon in a downturn; at the end of 2006, those reserves were actually a little larger, as a share of annual state expenditures, than before the recession at the start of this decade.

The problem is that recessions can have such large effects on state budgets that they tend to wipe out rainy day reserves and produce sizeable shortfalls. For example, states began this decade with reserves equaling 10.4 percent of annual expenditures, a quite substantial amount. But these reserves closed only about one-quarter of the state budget gaps that opened up through state fiscal year 2003. Moreover, if there is a recession in 2008, the effects on state and local revenues may be particularly large because declining home values are now causing property tax revenues to decline. (See Elizabeth McNichol and Iris Lav, “13 States Face Total Budget Shortfall of At Least $23 Billion in 2009; 11 Others Expect Budget Problems,” Center on Budget and Policy Priorities, December 18, 2007,

It is conceivable that federal fiscal relief could create a “moral hazard” problem if it filled most of or all of the state budget gaps that result during a recession, but it does not. The $20 billion in federal fiscal relief provided in 2003 closed only about 10 percent of the state budget shortfalls that emerged when the economy was weak in the early years of this decade.

[15] Ibid.

[16] “State Finances are Weakening,” National Conference of State Legislatures, December 10, 2007.

[17] The relief provided to states in the last recession, while helpful to states in addressing their fiscal difficulties, was not provided until mid-2003, two years after the recession hit bottom. That greatly reduced its value as stimulus. Relief to state governments can be very effective stimulus, if it is provided quickly once the need for stimulus is recognized.

[18] To the extent that major infrastructure projects are in states’ capital budgets, they would be less affected by a temporary downturn in economic activity than are expenditures in states’ operating budgets, because it is only the operating budgets that must be balanced each year.

[19] Depending upon their form, business tax incentives could put additional fiscal pressure on state governments whose business taxes are calculated using the federal corporate tax base. Accelerated (or bonus) depreciation or increased expensing of investment would erode those states’ corporate tax bases, thereby driving the states deeper into deficit and forcing them to raise other taxes or cut spending more deeply. A federal investment tax credit would not have that effect.

[20] Darrel Cohen and Jason Cummins, “A Retrospective Evaluation of the Effects of Temporary Partial Expensing,” Finance and Economics Discussion Series 2006-19, Federal Reserve Board, April 2006.

[21] GS Weekly, September 21, 2007.

[22] Congressional Budget Office, “Economic Stimulus: Evaluating Proposed Changes in Tax Policy,” January 2002,

[23] “Principles for Economic Stimulus,” Kent Conrad, Pete Domenici, Jim Nussle, and John Spratt, October 4, 2001.

[24] The expectation that stimulus measures will ultimately be paid for also could have a short-run benefit to the extent that evidence of fiscal responsibility keeps long-term interest rates lower than they otherwise would be and prevents “crowding out” of interest-sensitive investment and net export spending. Of course, economic weakness itself tends to be associated with lower-than-normal interest rates, so this effect may not be large.

[25] Auerbach and Feenberg, op. cit.

[26] The provisions comprise Title IV of the Trade and Globalization Act of 2007, H.R. 3920.