According to the National Bureau of Economic Research, the U.S. economy entered its tenth postwar recession early this year. At this point, and according to the Blue Chip forecasters, the recession seems to have been much milder than most, with the unemployment rate unlikely to go much above 6 percent. But in the early fall there were moments when the risk of a serious downturn was much greater, and even the relatively flat 2001 recession prompted renewed policy discussions about how to deal with recessions. Today, I offer my own views on the subject. For obvious reasons, I will avoid an evaluation of specific actions or pieces of legislation and just discuss a general strategy. My central point is that putting in place a sensible long-term macroeconomic policy strategy, covering both fiscal and monetary policy, is extremely important. The main importance of such a strategy involves the effects of the policy position on trend economic growth, not recessions or expansions. Business-cycle movements obviously have to be dealt with, but for the most part they can be addressed through the existing system of fiscal and monetary shock absorbers. At times, these shock absorbers may have to be augmented with discretionary antirecessionary fiscal actions, but these times are relatively rare. And the discretionary antirecessionary fiscal actions should meet stringent criteria to make sure that they do more good than harm. Sound Long-Run Policies The fundamental reason for establishing a sound long-run macroeconomic policy strategy is that the path of trend productivity growth is a nation's most crucial economic variable. It determines the long-run growth of living standards and is a main determinant of the affordability of entitlement spending programs, the level of household wealth, the value of the currency, and other economic magnitudes. Whatever policy does in response to short-term needs, that policy should be formulated in the context of a sound overall long-run strategy. The path of trend productivity growth has many underlying determinants--the vibrancy of technology, the strength of entrepreneurial motives, the flexibility of labor and capital markets, and the effectiveness of regulatory policies, to name a few. But there are also two vital macroeconomic policy determinants. One, involving monetary policy, is a credible commitment to stable prices in the long run. The second, involving fiscal policy, concerns the share of total output devoted to national saving. Output that is not devoted to saving will be consumed, and that's that. Output that is saved, or reserved from consumption, will be devoted to capital investment, which equips workers with new capital, enhances productivity, and raises living standards. In recent years, the United States has been the fortunate beneficiary of a productivity growth spurt. Both monetary and fiscal policy have played an important role. The commitment to and realization of stable prices have surely been critical background forces, facilitating long-term planning and a more efficient allocation of resources. On the fiscal side, the role of saving and investment has been important as well. Production-function decompositions of this growth spurt suggest that a good 30 to 40 percent of it has been due to so-called capital-deepening productivity change, reflecting increases in the economy's capital-labor ratio. Fiscal policy is relevant because it helps determine the national saving that finances this capital investment. National saving can be generated privately, by households and business, or by government. Efforts to promote private saving certainly have merit in theory, but the bald empirical fact for the United States is that the government has introduced a number of tax provisions to increase private saving over the past twenty years, and almost all of them have been correlated with a reduction in the overall rate of private saving. These measures could have been effective at the margin, but they have not had a large enough effect to counter the general downward trend in private saving. As an added caution, from a standpoint of national saving, it makes no sense to introduce costly private saving incentives that then reduce government saving more than they raise private saving. What has seemed to raise national saving is government fiscal policy, the budget surpluses. Government saving, or a surplus, leads to a reduction in outstanding government debt to which wealth-holders, for the most part, seem to respond by devoting the funds to private credit markets. This response, in turn, finances capital investment. Much of this capital investment is likely to be in the United States, directly influencing output and productivity. But even if the saving flows abroad, either through the buying back of foreign-owned capital or the purchase of new capital in foreign countries, the United States gets the returns on this investment, which raise income growth and future consumption. The only instance in which government surpluses do not raise national saving is the so-called Ricardian equivalence scenario. In this scenario, households learn that the government plans to run a surplus, and they reduce their own saving by their proportionate share of this surplus. In general, economists find that if Ricardian offsets exist at all, they are a modest share of the relevant government surpluses. At a more casual level, I have, for many years surveyed my non-economist friends to see how they alter their own consumption behavior in response to national fiscal policy. The usual response is a blank stare: If these non-economists have even imagined there to be a connection between government deficits and their own saving, they have no idea how much to respond and in what direction. Fiscal Policy What, then, is the long-run fiscal policy that leads to a vibrant economy with a high rate of productivity growth? In general, it implies nudging the economy in the direction of raising the overall rate of national saving. It is theoretically possible for an economy to save so much that the long-term path of consumption per capita is actually lowered, but generally economists believe that the U.S. economy is well short of that saturation point. For the present circumstances, I am proceeding on the assumption that more saving is better. I start by decomposing the federal budget into two components, the cash surplus of the trust funds for social security and some other entitlement programs, and the net surplus for all other activities of government, which I will call the general budget surplus. For the general budget surplus, let me repeat a very simple rule, similar to that enunciated by both the Committee for Economic Development and Milton Friedman fifty years ago. The rule is that the general budget surplus should average zero over the long run. This general budget should not be exactly balanced each year, but there should be long-run balance. In the event of a recession, for example, it is natural, and actually healthy, for revenues to fall, spending to rise, and the general budget to move into deficit. This deficit helps prop up the economy in the recession, which is why these budgetary movements are called automatic fiscal stabilizers. But though deficits are excusable in a recession, they are not desirable in general. If the following period sees an expansion, the general budget should move into surplus, and over time the general budget surplus should average zero. Indeed, the long-term value of higher national saving gives an argument for aiming at a positive general budget surplus on average, but I will not push that argument here because it is most likely a political nonstarter. For the surpluses of the entitlement program trust funds, covering social security, Medicare, and the retirement benefits of government employees, one would follow a different rule. These trust funds are set up to finance long-term entitlement expenditures. To illustrate, for social security, the largest of these entitlement programs, revenue from worker payroll taxes go into a trust fund and, along with any earnings on asset holdings, are used to pay retiree benefits. These trust funds should then be evaluated on an actuarial basis--whether present and future anticipated inflows are adequate to pay anticipated present and future program benefits. If not, corrective steps should be taken. Again using social security as an example, inflows currently exceed outflows (a cash surplus), but forecast inflows do not cover forecast outflows (an actuarial deficit). Even though the fund has a cash surplus today, some combination of tax increases or long-term cuts in benefits is needed to maintain these surpluses into the future and to restore actuarial soundness. This situation is not atypical for countries like the United States with an aging population. For a country with an aging population, the combination of a general budget surplus that is zero on average and trust funds in cash surplus will normally lead to surpluses in the overall, or unified, budget. These surpluses can be rationalized as a technique to build up the nation's future capital stock. Since the population is aging, a relatively small cohort of workers will soon have to pay benefits for a relatively large cohort of retirees, and the idea is at least to give this small cohort of workers a larger capital stock to work with. Having this type of a budget policy accomplishes three objectives simultaneously: The surpluses provide for capital accumulation. They do that to a greater extent the more the country's population is aging. Automatic fiscal stabilizers remain to deal with recessions. The discussion to this point has been entirely in terms of the federal budget. In principle, state governments could follow the same countercyclical fiscal policy, except for one problem. States typically operate under legislative or constitutional limits on their deficits. The only way states could run deficits in recessions is if they had saved a store of assets in earlier booms. This presaving strategy sounds nice in principle and is a good idea, but very few states actually follow it to a sufficient degree. So, as state budget situations worsen because of the state automatic fiscal stabilizers, the unfortunate but likely result is that states will lack the rainy day resources to run deficits and will need to make budget changes that neutralize their own automatic stabilizers. Monetary Policy The automatic fiscal stabilizers help in stabilizing the economy over the business cycle, but the main burden of fighting recessions is left to the central bank. Traditionally there have been two reasons for this: Monetary policy responds to shocks much more quickly than fiscal policy. It is generally believed that, if an economy is open to foreign trade and capital flows and has flexible exchange rates, monetary policy becomes more effective in fighting cycles whereas fiscal policy becomes less effective. To be sure, one can raise questions about the second reason--the argument hinges on the fact that the value of the domestic currency is positively correlated with interest rates, which in recent years has often not seemed to be the case for the United States. But despite any qualms about the second reason, the first reason seems solid for just about any known real world economy. For a long time now, central banks have been trying to stabilize economies, in an art form known as "leaning against the wind." John Taylor has developed an equation that provides one statistical description of how this practice might be followed. Under what is known as the Taylor rule, there is a certain equilibrium position for some short-term interest rate. The central bank uses open market operations to raise the short-term rate above this equilibrium in response to threats of inflation and to lower the rate below equilibrium in response to threats of unemployment. I won't go into details, but if the central bank is willing to announce an inflation target, the policy responses implicit in a Taylor rule could be similar to those implicit in forward-looking, flexible inflation targeting, a policy strategy now used successfully in controlling inflation by about ten leading economies around the world (twenty, if one counts the European Central Bank, which practices a form of inflation targeting). Determining how much central banks should raise short-term interest rates in response to inflation, or lower them in response to recession, or even in response to subpar growth, becomes an empirical matter. However much they do, we have another stabilizer for the economy. As demand growth heats up, the central bank will raise short-term rates to try to limit the boom and the resulting inflation. As recessions threaten, the central bank will lower short-term rates to try to limit the downturn. Discretionary Fiscal Policy? With this long-run strategy in place, and with at least two shock absorbers--one monetary, one fiscal--we are now ready for the sixty-four dollar question. In the average recession, should there be additional discretionary fiscal actions? For such actions to be effective, they must satisfy three conditions: They should stimulate spending. They should be instituted quickly. They should be reversed quickly or automatically. The reason speed and reversibility are important is that most recessions are not very long, averaging about twelve months in the United States. If some expansionary action comes into effect late in a recession, say in the twelfth month, it is likely to add to the vigor of the next expansion and hence increase business-cycle volatility. Reversibility is also important because of the overall budget rule discussed earlier. If the general budget were already in approximate balance, any continuing expansionary fiscal action taken on top of that would generate normal deficits, not normal budget balance. There is a limited set of fiscal actions that satisfy all these conditions. Many of the commonly discussed proposals fall short on at least one ground. By way of illustration, I will discuss a few that have been mentioned in recent debates. Extend unemployment insurance. Unemployment insurance benefits run out after twenty-six weeks in the United States. In general having a time limit to control caseloads is helpful--the basic idea of unemployment insurance is to provide temporary, not permanent, support to unemployed workers. But in recessions, when unemployed workers have much more difficulty in getting new jobs, it may make sense for policy to be more lenient by extending time limits. Such extensions have been enacted in nearly all postwar U.S. recessions. The extensions would provide benefits to unemployed workers who might have trouble maintaining their pre-unemployment standard of living, implying that the extensions should stimulate consumption spending. The extensions, say to thirty-nine weeks, could be made for a defined period, such as a year. Even if the period exceeded a year, most of the budget dollars would be reversed as workers would naturally be better able to get jobs and would not be remaining on the unemployment rolls. This is one action that seems to satisfy all three conditions. Accelerate scheduled cuts in tax rates. If future tax-rate cuts are already enacted into law, putting them into effect sooner than planned could be done quickly. Presumably the relevant political compromises have already been made. As the acceleration would not damage the long-run budget position, the reversibility condition is satisfied. But that property could lessen the effect of the rate cuts. It has been found that consumers spend relatively little out of one-shot increases in their income. They spend much more when tax cuts are viewed as permanent rather than one-time events. Since rational, forward-looking consumers should view accelerated rate cuts as one-shot increases in income, the effect of these cuts on consumption could be limited. On the other hand, if consumers are not forward-looking but wait for their tax cuts before they begin spending, the accelerated rate cuts could stimulate a normal amount of consumer spending. In this case accelerated rate cuts could satisfy all three conditions. If one changed the tax policy expectations, the overall verdict is, if anything, less favorable. Suppose households did not expect proposed future rate cuts to become law. In such cases, moving up rate cuts might have a stronger perceived effect on permanent income and consumption and more clearly satisfy the impact test. But the same action might take the general budget out of long-term balance, at least as far as bond markets are concerned, and would fail the reversibility test. In this latter case, the accelerated tax cuts satisfy only two conditions. Temporary cuts in payroll taxes. One could also respond to recessions by cutting payroll taxes for a limited period. The cuts could be made quickly, satisfying the first requirement. But as with income taxes, if rates are cut only for a short time, the funds will appear to be temporary to consumers, in which case their effect on consumption may not be very large. Another problem with changing payroll taxes involves the fact that virtually all of them are now devoted to entitlement trust funds such as social security. The social security trust fund is already underfunded in a long-term actuarial sense, and payroll-tax revenues could not be removed from the fund, even for a short period, without raising the long-term actuarial deficit. Moreover, one would not want to start the tradition of using a long-term program to fund short-term needs--soon no long-term program would exist. Thus, if payroll taxes were to be cut, the fiscal flows would have to be kept separate from the social security system. Income taxes would need to be cut by an amount that workers would otherwise have paid into the fund as payroll taxes, without actually taking money from the fund itself. Such logistical complications could be difficult to work out and could lead to delays in enactment or implementation. Temporary cuts in sales taxes. A fast-acting, short-term measure with a powerful spending impact is a temporary cut in sales taxes. Whereas temporary income-tax cuts can be viewed as making a one-shot change in household disposable income, temporary sales-tax cuts bring about a temporary reduction in the prices of goods that consumers buy. The switch from a temporary income change to a temporary price change makes all the difference. The government is offering a sale on consumption goods and consumers should respond by raising consumption. The period of reduction should be kept reasonably short, to encourage households to take advantage of the favorable terms. It appears that this measure might satisfy all three tests. The only problem with temporary sales-tax cuts at the federal level is that the federal government does not impose any sales taxes. Most states do. Hence, for this measure to be effective, the federal government must give grants to those state governments that cut their sales taxes, roughly in the amount of the revenue lost to states from the cuts. Other arrangements might need to be made for the few states that do not impose sales taxes. Our supposedly quick program just got a bit more complicated, and one can easily imagine various states arguing about the revenue flows in the bill, holding up passage until the cyclical need has passed. Temporary inducements for investment. Following the same principles, the federal government, which does levy corporate taxes, could offer a sale on investment goods. Until 1986, the normal way of doing this was with a temporary investment credit. But the new tax laws make such a measure much more difficult. What can be done instead is to offer a temporary acceleration of depreciation allowances. As with temporary sales-tax reductions, making the period reasonably short could accelerate investment expenditures to the time when aggregate demand is weakest. A pro and a con are associated with focusing on investment instead of consumption. The advantageous feature is that investment represents capital deepening--the whole goal of keeping the long-run general government budget in balance is to encourage an investment-oriented economy. The con is that investment is already a more volatile spending component than consumption. If the temporary investment inducements are not timed to coincide with the cyclical dip in investment, they could exacerbate investment cycles and instability. But under some circumstances the investment inducements could pass all three tests. Conclusion The important point is to establish the right long-run policy framework, with general budget fiscal policy aimed at establishing budget balance on average over the cycle and entitlement trust fund policy aimed at making these programs actuarially sound over the long run. Monetary policy should be aimed at long-run price stability, with short-term flexibility in response to demand conditions. The fiscal measures and the monetary commitment to price stability stimulate investment, productivity, and technological change over the long term. The flexible monetary policy and the automatic fiscal stabilizers provide the economy with shock absorbers to smooth out business cycles. The unanswered question then involves the role of discretionary fiscal changes-- are they a valuable addition to this macroeconomic strategy? To be valuable, the measures should stimulate spending, take effect quickly, and be easily reversible. A number of temporary tax changes might do the job, but if they do not stimulate spending, their enactment is delayed, or if they are not easily reversible, they might do little more than to undermine the long-run fiscal strategy. The outcome ultimately depends on the details of the program, the speed of enactment, and the ease of reversibility.

Return to top 2002 Speeches