Mr. Chairman, Members of the Committee,
Thank you for the opportunity to testify on this important topic. In my testimony today, I would like to make five points.
- Nominal and real interest rates are likely to remain low for a long time to come. Indeed, nominal interest rates are forecast to be lower than the growth rate of nominal GDP for the next 20 years. This is not an absolute certainty, however, and one must take this uncertainty into account.
- Low real rates have three implications for fiscal policy. Fiscal costs are lower: The cost of debt is currently negative. Primary deficits now and in the near term require smaller primary surpluses in the longer term. Fiscal risks are accordingly lower: There is little chance of a market-based US debt crisis in the foreseeable future.
- Low nominal rates put sharp limits on the use of monetary policy. The most the Federal Reserve can do to stimulate the economy is to decrease nominal rates to zero or close to zero. Once at the lower bound, monetary policy cannot help much. But fiscal policy can.
- Lower fiscal costs and higher benefits imply a larger role for fiscal policy as a macro stabilization device. Put another way, the trade-off between debt stabilization and output stabilization has shifted in favor of output stabilization.
How these conclusions apply to the United States:
First, deficits, running at more than 5 percent of GDP, are large. Unless they are used to finance an ambitious and credible public investment plan, they should be decreased. Decreasing them too fast would be risky however, as there is little room for the Fed to decrease interest rates. The reduction in the deficit should be contingent on the strength of private demand. Following this strategy might lead to further increases in the ratio of debt to GDP, but this is an acceptable risk.
Second, if a recession materialized, monetary policy would likely be constrained, making it essential to use fiscal policy. Automatic stabilizers, as they currently exist, are too weak. Better ones should be designed soon.
Let me elaborate briefly on each of these points:
Interest rates are likely to remain low for a long time.
This is certainly the belief embodied in US bond prices. The yield curve shows 30-year rates below 3 percent, very likely lower than the nominal output growth rate over the same period. Option prices indicate that investors put little probability on a sharp increase in rates in the future. For example, investors put the implicit probability that the short rate will exceed 4 percent in five years at less than 10 percent. And low rates reflect a worldwide phenomenon: Rates are low in most advanced economies, indeed lower than in the United States.
History tells us that markets do not always get it right, and that they tend to react too late and then by too much. Thus, it is important to look for what factors may have triggered these low rates, and whether they can be expected to remain or to change in the future.
What is clear is that the low rates reflect more than the lasting effects of the financial crisis. Their decline is a long-standing trend, starting in the mid-1980s. It is fair to say that, while many factors have been identified as potential causes, ranging from an aging population to precautionary saving, to lower growth, to a higher demand for safe assets, we are still uncertain as to the role of each one. What can be said, however, with more confidence is that none of these factors appears likely to reverse any time soon.
Finally, when it comes to fiscal policy implications, even if rates were to increase substantially in the future, the government can largely lock in the low rates by issuing long maturity bonds. The bond market looks sufficiently deep up to long maturities to absorb a substantial increase in the supply of such bonds.
Low rates have three main implications for fiscal policy.
The first is, rather trivially, that they decrease the fiscal costs of debt. The proposition can be put in an extreme, provocative, but useful way. When the interest rate is less than the growth rate, an increase in the deficit today, due to, say, higher spending or lower taxes, does not need to be offset by higher taxes or lower spending in the future: While debt will increase at the rate of interest, the economy will increase at the rate of growth, leading to a decrease in the ratio of debt to GDP over time. The proposition comes with strong caveats, as one cannot be sure that the interest rate will be less than the growth rate in the future, but it is nevertheless a useful counterweight to the proposition that debt has very high fiscal costs.
The second implication, which is closely related, is that low rates decrease the fiscal risks of debt. Fiscal crises typically come from the perception by investors that the primary surpluses that need to be generated to stabilize the ratio of debt to GDP are becoming so large as to be politically unfeasible, forcing the government to default. When the interest rate is less than the growth rate, however, stabilization of the debt-to-GDP ratio does not require a primary surplus but instead allows for a small primary deficit (based on current numbers, the Federal government can run a primary deficit of 1 to 2 percent and maintain a roughly constant debt-to-GDP ratio). This makes any debt crisis very unlikely.
The third implication is less obvious but in some ways more important. It is that debt has not only low fiscal costs but also low economic (or, equivalently, welfare) costs. All other things being equal, higher public debt leads to less capital accumulation and thus to lower future output. This cost, however, depends on the rate of return to capital. And the signal sent by the low safe rate, which we can think of as reflecting the risk-adjusted rate of return to capital, is that, risk adjusted, the rate of return to capital is low. Thus, crowding out of capital, to the extent that it takes place, is not very costly. This conclusion again comes with plenty of caveats (having to do with uncertainty, measurement issues, and so on), but it is again a strong counterargument to those who argue that high debt will kill growth.
The fact that low nominal rates decrease the room of maneuver of monetary policy is now well understood. Nominal rates cannot go far below zero before triggering a shift to cash. This constraint used to be called the zero lower bound. The experience of foreign central banks has shown that the nominal rate can be slightly negative, leading the constraint to now be called the “effective lower bound.”
Combined with low inflation, this bound implies that real rates cannot go very negative. (The issue would be less relevant, were the inflation rate higher, but this is another discussion.) The last 10 years have shown that the Fed has tools other than the policy rate to boost activity, but it is also clear that it does not have enough room today to react as it would like to a serious decrease in activity, whether this decrease is accidental or is the result of a fiscal contraction. A back-of-the-envelope computation suggests that a fiscal contraction of say 1 percent of GDP might force the Fed to return the funds rate down close to the zero lower bound.
Lower costs of debt on the one hand, and limits on monetary policy on the other, change the trade-off between debt and output stabilization. This has two straightforward implications for fiscal policy.
Less emphasis on fiscal consolidation: It may still be that the desirable long run level of public debt is much lower than the current level. It may even be negative. Future fiscal and off-balance sheet obligations, intergenerational distributional preferences, and the cost of fighting global warming and its implications are all relevant here, and I do not pretend to have an answer. But, even if the desirable level of debt is lower than the current level, it is still the case that low rates imply that fiscal consolidation is both less urgent and potentially more costly in terms of reduced output.
More willingness to use fiscal policy to fight a recession: Were aggregate demand to slow down and a recession become likely, fiscal policy should be used more aggressively than in previous recessions. The cost of higher debt from such an aggressive response is likely to be much smaller than the output cost from a more limited response.
- Going from these general principles to specific recommendations for US fiscal policy at this juncture is, as always, more difficult, and there is room for disagreement. Nonetheless, these are my recommendations.
Current deficits are running at more than 5 percent. Current primary deficits are running at close to 3.5 percent, thus above the 1 to 2 percent level that would stabilize the debt-to-GDP ratio. Should they be decreased or increased, and if so, at what rate?
The case for increasing the deficits at present is not as irresponsible as it may first sound. Suppose that deficits were increased by, say, 1 percent of GDP, and that to avoid overheating, the Fed increased interest rates by 1 to 2 percent, thus getting further away from the lower bound. The cost of increased debt might be offset by the benefits of increasing the room for action by the Fed, in terms of insurance against the next recession.
That case however is sufficiently uncertain that I believe that the current goal should still be to decrease primary deficits. If so, the fiscal strategy should be to decrease them at a speed that allows the Fed to offset the adverse effects of consolidation through lower interest rates. The speed therefore should be contingent on the strength of private demand.
An increase in private demand, if it were to happen, would give more room for consolidation; persistent low demand may instead require maintaining large deficits for some time. In that case, the large deficits might lead to a further increase in the debt-to-GDP ratio; the fiscal costs of such an increase are likely to be smaller however than the output costs that would result from a faster consolidation.
To the extent that deficits remain large, they should ideally be used for capital spending rather than current spending. Gross government capital spending has decreased from 3.9 percent of GDP in 2000 to 3.2 percent today. Over the same period, net government capital spending has decreased from a small 1 percent of GDP to an even smaller 0.5 percent. This is a worrisome evolution, especially in the light of increased demands to slow global warming and deal with some of its worst implications.
In short, judicious use of deficits as a way of simultaneously sustaining demand and output in the short run and financing public investment and increasing output in the long run appears today to be the best strategy.
Finally, given the limited room for monetary policy to help, fiscal policy must be ready to fight the next recession when it comes. US automatic stabilizers are weak. Discretionary policy measures take too long to pass and to implement. It is thus essential and urgent to put in place “semi-automatic stabilizers,” automatic changes in the tax system or in spending, triggered by the evolution of unemployment or output. Several proposals are on the table. They should be examined, and some of them should be adopted, sooner rather than later.