Lawrence Summers Testifies Before Joint Economic Committee

Fiscal Stimulus Issues
Testimony before the Joint Economic Committee
January 16, 2008
Washington, DC

Lawrence H. Summers
Charles Eliot University Professor
John F. Kennedy School of Government
Harvard University
I am grateful for the opportunity to testify before this committee at this important juncture. I have submitted for the record a speech I delivered recently on the current state of the economy and the broad range of policy questions it presents. Here I address the main issues at stake in the debate over fiscal stimulus for the American economy and provide my views as of this moment. The best policy response may change as we receive new economic data and our understanding of the current, highly volatile economic situation improves.

1. Is fiscal stimulus desirable at present?

Yes. Following recent economic reports particularly the last employment report and yesterday’s retail sales data, my judgment, like that of many economists, is that a recession is more likely than not. At this point this is or is very close to being a consensus judgment. Even if there is not an officially defined recession, there is almost certain to be a significant slowdown in the economy that will feel like a recession in many parts of the country and to many businesses and families.
Fiscal policy measures that succeed in increasing spending will mitigate the severity of the coming downturn and accelerate the eventual recovery. Crucially they will also provide insurance against the possibility of a scenario unfolding like that in Japan in the 1990s where an economic downturn becomes very severe and lengthy because a vicious cycle starts with credit problems hurting the real economy, in turn exacerbating the credit problems with the cycle continuing.
Over the next two years the difference between economic performance with and without fiscal stimulus is likely to be several hundred thousand jobs and a loss in the range of $1000 for the average family. If fiscal stimulus is successful in preempting a severe recession that would otherwise occur, the gains would be far larger.

2. Why not rely on monetary policy to stimulate the economy and focus fiscal policy on longer term issues?

As Chairman Bernanke recognized in his most recent speech, monetary policy has an essential role to play in maintaining demand and growth as well as in combating financial instability. In the current context however it is best complemented by fiscal policy for a variety of reasons: (i) in normal times fiscal policy is faster acting than monetary policy, and given the financial problems it may be even more true today. (ii) proper fiscal policies can target the innocent victims of recession and can directly promote job creation (iii) full reliance on monetary policy could easily mean lowering interest rates to levels that would be problematic for the dollar, commodity prices, future asset bubbles and moral hazard (iv) in a situation where policy impacts are uncertain it is most prudent to rely on a diversified set of stimulus measures.

3. How great is the risk of overheating the economy and causing inflation? Should a decision on fiscal stimulus await definitive evidence that the economy is in recession?

The balance of risks is now on the side of recession rather than inflation. Inflation measured by personal consumption expenditures excluding food and energy was 1.9 percent over the last year. Measures of inflation expectations as inferred from Treasury indexed bonds are close to their lowest point in the last two years. Moreover, in a climate of great uncertainty about workers’ jobs and firms’ profit margins inflation pressures are more likely to diminish than increase. Increases in inflation that have been observed recently reflect to a significant extent the impact of developments in oil as well as other commodity markets as well as declines in the dollar. Even if they are not reversed, these markets are unlikely have as large an inflationary impact in the future as in the recent past.
The major problem with past stimulus efforts is that the stimulus has come too late. If stimulus is to be timely, it should be delivered promptly. By the time it is conclusively established that a recession has started, policy is likely to have been substantially delayed from what would have been optimal.
There is sufficient weakness in the economy to justify stimulus legislation now with provision for rapid implementation. As discussed below, consideration should be given to providing for a second tranche of stimulus that would be implemented if/when clearer evidence of economic weakness appears.

4. How large should a stimulus package be?

I have previously advocated stimulus in the range of $50-$75 billion. Given recent data, I now believe that it would be appropriate to enact a program of this magnitude as soon as possible and to make provision--perhaps with a contingent trigger for a second tranche of about the same magnitude.
Sizing a stimulus package cannot be reduced to hard science. A program of $50-$-75 billion would represent .35%-.55% of GDP and would run very little danger of overheating the economy on any plausible scenario. If delivered in the second and third quarters of 2008 it could have a material impact on consumers and on confidence more generally. A larger program particularly if it contained measures that would have their impact only with a delay would on current evidence risk an imprudent pattern of expanding the budget deficit at a future time when growth had been restored.
However, policy has been behind the curve for some months now. So it would be appropriate to design further stimulus that could take effect without the need for new legislation and debate if the economic situation deteriorated either as Congressional debates continue or after legislation takes effect. The two tranche approach could deliver stimulus in the range of 1% of GDP if the situation warranted. One reasonable trigger would be changes over a quarter in payroll employment as suggested by Martin Feldstein. An alternative or supplement would give the Executive Branch the right to trigger the second tranche of stimulus based on their assessment of economic conditions.

5. What should comprise a stimulus package?

As with any potent medicine, stimulus if misadministered could do more harm than good by increasing instability and creating long run problems.
A stimulus program should be timely, targeted and temporary.
Timely stimulus requires both that Congress and the President act quickly and that measures be chosen which can be implemented rapidly and then which will have their ultimate impact on spending in short order. This puts a premium on simple measures that work through existing modalities such as adjustment of withholding schedules or tax refunds or enhancements of benefits. It calls into question the wisdom of designing new programs or using approaches where Federal spending is not injected fairly directly into the economy but is instead only transferred to other units of government that historically have spent out new funds only gradually.
Targeted stimulus requires that funds be channeled where they will be spent rapidly and where they will reach those most in need. This also argues for use of simple changes in withholding schedules, or tax refunds as well as for change sin benefit formulas. In general targeting in both the sense of assuring maximum spending and fairness are likely both to be achieved by measures that target those with low incomes and whose incomes have sharply declined.
Temporary stimulus is necessary if stimulus is not to raise questions about the country’s long run fiscal position. If stimulus were not credibly temporary, it would likely raise long term interest rates and increase capital costs offsetting its positive impact. Moreover if stimulus is not temporary, the risks that it will continue even after the economy recovers and lead to inflation or very high interest rates is greatly increased. Stimulus should be designed so that its proximate impact on consumer or government spending is all felt within a year of enactment and in any event by the end of the first quarter of 2009. It is important also that no measures be enacted on a temporary basis that will generate overwhelming political pressures for their extension if fiscal credibility is to be maintained.
On the tax side these considerations suggest the desirability of across the board equal tax cuts or refunds for all taxfilers Measures which reduce taxes in proportion to taxes currently paid or that disproportionately favor upper income taxpayers or recipients of capital income are likely to be far less effective as such taxpayers spend much less of new income than low and moderate income taxpayers. Measures which commit today to reduce future taxes relative to current law are likely to be counterproductive because of the fiscal doubts they raise and because they do not provide liquidity now at the moment when consumers are facing the need to cutback spending.
My reading of the evidence suggests that the case for business tax cuts is not compelling. The principle inhibitor of business investment right now is lack of market demand not the cost of capital. And the experience with the 2001 stimulus program is not very encouraging with respect to the efficacy of business incentives as stimulus. On the other hand a temporary investment tax credit or accelerated depreciation scheme might pull some investment forward from future years into 2008. To maximize the bang for the buck such a program should be incremental and apply only to investment above some benchmark such as 2/3 of previous investment or depreciation.
On the spending side the measures most likely to be effective are temporary increases in benefits perhaps for the long term unemployed and food stamp recipients. Such increases can be implemented quickly and go to people who will spend them fast. The more detailed efforts are made to achieve specific programmatic objectives through increased spending, the greater the risk of delays. A high burden of proof should be put on any new spending program proposed as a stimulus measure to demonstrate that the spending will be rapid.

6. Should stimulus be paid for within a given budget window?

Fiscal stimulus to an economy in recession operates by increasing demand in an economy that is constrained by lack of demand. If it is paid for contemporaneously, its point is largely lost as there is no net stimulus to demand because money injected in one area is withdrawn in another.
As long as a fiscal stimulus program is temporary and does not create expectations of future spending or tax cuts it does not make a large economic difference whether or not it is offset by specific future fiscal actions. Including offsets in a five or a ten year window would magnify the impact of fiscal stimulus a little bit by reducing any adverse impact on capital costs because it would avoid any increases in long run debt levels. And the need for offsets might operate to prevent fiscal stimulus from being extended or allowed to grow excessively.
On the other hand, if it delayed the process of coming to a conclusion on fiscal stimulus or generated so much disagreement that timely legislation could not be enacted a requirement of offsets would have serious adverse consequences. A judgment on offsets is ultimately political--depending on the economic advantage of the extra discipline offsets bring relative to the disadvantage of the extra complexity and delay that their negotiation would require.


John F. Kennedy School of Government 79 John F. Kennedy Street
Cambridge, MA 02138
617-495-1100 Get Directions Visit Contact Page