How Much Would the American Rescue Plan Overshoot the Output Gap?

The Congressional Budget Office (CBO) projected on Monday that the nation’s output gap – the difference between actual economic activity and potential output in a normal economy – would be $380 billion for the rest of 2021. This gap will total roughly $300 billion in the last three quarters of 2021 and nearly $700 billion through 2023, the period over which most future relief would take place. Assuming CBO's estimates are correct, President Biden’s $1.9 trillion American Rescue Plan would likely be enough to close the output gap two to three times over. This overshoot could be beneficial, but could also pose risks to the economy and the fiscal outlook.

What Would the American Rescue Plan Mean for the Output Gap?

The theoretical effect of the American Rescue Plan on the economy depends on the economic multiplier associated with the new programs, but in almost any circumstance it would substantially overshoot the output gap as estimated by CBO. With a multiplier of 0.5x, for example, the plan would close 135 percent of the output gap.1 With a 1.5x multiplier, it would close the output gap 4 times over.

Based on a recent analysis of the plan from Wendy Edelberg and Louise Sheiner of the Brookings Institution, the American Rescue Plan would theoretically boost output by about $1.5 trillion starting in the second quarter of this year, closing about 225 percent of the output gap through 2023.2 This estimate appears to be reduced by the fact that the economy would be performing in excess of its capacity, leading to some additional savings and inflation.3 Multipliers from their October 2020 analysis suggest that if the output gap were large enough, the plan could produce closer to $1.9 trillion of additional output and thus close about 275 percent of the output gap.

The output gap could differ from CBO's projections. Many forecasts and experts suggest the economy will grow faster this year than CBO estimates. A one percentage point increase in Gross Domestic Product (GDP) growth would reduce the output gap to less than $200 billion, in which case the American Rescue Plan would be large enough to close eight to ten times the output gap based on the Edelberg and Sheiner numbers. On the other hand, many have argued that CBO is underestimating full employment and potential GDP. If potential GDP were 1 percent larger than CBO's estimate, the output gap would total $1.3 trillion through 2023 and the America Rescue Plan would close 115 to 145 percent of the output gap.4

What Happens if We Overshoot?

Though many economists and experts such as Federal Reserve Chairman Jerome Powell have argued it would be better to overshoot with fiscal and monetary support than undershoot, that does not mean an overshoot of any magnitude is desirable.

While the economy can operate above its long-term potential for periods of time, it cannot do so indefinitely or sustainably. It is therefore important to understand what might happen if policymakers spend substantially more than what is necessary to close the output gap.

One possibility is that the excess funds are economically ineffective, adding to the debt without improving the economy. Thought of another way, overfilling the fiscal gap could substantially reduce economic multipliers.5 Based on estimates from CBO, adding nearly $2 trillion to the debt would shrink the size of the economy by about 0.3 percent ($100 billion) by the end of the decade while increasing annual debt service payments by roughly $40 billion in that year (and more in future years).6 These costs are probably a worthwhile consequence of addressing an economic crisis and restoring the economy to full employment but harder to justify for spending that has little or no economic impact.

A second possibility is that excess funds could lead to higher inflation, with producers responding to higher demand by increasing prices once it is no longer possible to easily increase production. In some ways, higher inflation could be helpful – it could erode outstanding household and business debt (including pandemic-related debt), lower real interest rates set by the Federal Reserve,7 and help the Fed to reset expectations toward its new flexible inflation targeting regime. On the other hand, higher inflation could also diminish the effectiveness of the fiscal stimulus, erode the value of savings (including the over $2 trillion of personal savings accumulated because of the pandemic), increase the cost of living for many households who could not easily afford it, or, in the worst case, lead to persistently high inflation and all the consequences that come with it.

A third possibility is that excessive stimulus could cause misallocations in the economy. Higher demand amidst a pandemic could lead to increased consumption and production of goods and services that are of less value in normal times. To the extent that firms and households make long-term investments in response to near-term demand, this could cause modest macroeconomic damage in the long term as well as diminishing welfare gains in the near term.

A fourth possibility is that excessive stimulus could temporarily boost economic activity far above its sustainable potential, leading to an economic cliff or crash as the stimulus fades. In their estimates, Edelberg and Sheiner explicitly assume what they describe as a “soft landing” for the economy; even in this scenario, there appears to be virtually no economic growth in 2022, which suggests unemployment would rise over that period.8 The authors warn of the possibility of a sharper and more painful contraction (a "hard landing") when stimulus funds run out.

To prevent this cliff, lawmakers may have to enact more and more stimulus. In that case, the economy may become dependent on ever rising deficits just to maintain sufficient demand. Aside from any inflationary pressures they might cause, these large deficits would put debt on an even more unsustainable path, substantially boost interest payments, and impose substantial damage on the broader economy. Even as ongoing stimulus might continue to keep the output gap at bay, it would increasingly weaken potential GDP. In an extreme case, an additional $2 trillion per year of borrowing could lift debt to 175 percent of GDP within a decade, boost interest payments by up to $500 billion per year by decade's end,9 and slow economic growth by 0.3 percent per year.

To be sure, there are also potential benefits to overshooting. Providing more stimulus than needed can insure against the risk that multipliers are lower than expected, underlying economic conditions are worse than expected, or potential GDP is higher than believed. Excessive fiscal stimulus in the near term could also prevent long-term economic damage from the recession, for example, by preventing viable business closures or stopping long-term unemployed workers from dropping out of the labor market (known as hysteresis). Since it is impossible to perfectly target spending toward the precise needs of society, it may also take more funds to address the specific “bottom-up” weaknesses in the economy than to close the top-down output gap.

Still, with $4 trillion of fiscal support already enacted, it is not clear whether another $1.9 trillion is needed from a bottom-up or top-down approach. While recent data suggest further fiscal support is needed, the package currently under discussion would likely be an overshoot.


1 We assume the plan would be enacted in March and have no substantial macreconomic impact until the second quarter of 2021.

2 Though we have not run our own analysis, applying our December multipliers to the plan would yield a similar result.

3 The authors note that they "attenuate the MPCs to account for recipients’ responses to temporary increases in prices that would result from the increase in scarcity of goods and services and because the saving rate has increased sharply during the pandemic, making households less likely to be liquidity constrained."

4 Edelberg and Sheiner assume potential GDP would be halfway between CBO's January 2020 and July 2020 forecast. In that case, the output gap would total $525 billion through 2023 and the American Rescue Plan would be large enough to close 285 percent of the gap. These output gaps differ slightly than the ones we published yesterday because those discuss the output gap starting in February. These figures assume the bill's effects begin in April (the start of the second quarter).

5 The Congressional Budget Office has estimated the $2.64 trillion of net COVID relief enacted earlier in 2020 would reduce real GDP by 0.4 percent by 2030.

6 Fiscal support would have little effect on the macro economy if most new spending was saved rather than spend and/or of spent funds mainly pushed up prices rather than increased the amount of goods and services bought and produced.

7 The real interest rate is the equal to the nominal interest rate minus inflation. For example, if the federal funds rate is 3 percent and inflation is 2 percent, the real interest rate is 1 percent. During economic downturns, it often makes sense for the Federal Reserve to set negative real interest rates. Because it is difficult to set nominal rates below zero, however, the negative real interest rate cannot go lower than the inflation rate. For example, if the inflation rate is 2 percent then the real interest rate can go as low as -2 percent. But if the inflation rate is 3%, the real interest rate can fall to as low as -3 percent.

8 Edelberg and Sheiner state that real GDP would contract slightly in the middle of 2022. If GDP is flat, it likely means there is either job losses or losses in per-person income. Since productivity would be expected to rise during that time, job losses would be more likely to happen, though either happening would be troubling.

9 The interest estimate comes from both debt service on that additional debt and the assumption that interest rises by 0.025 percent for every 1 percentage point increase in debt, consistent with CBO's assumptions.