Author: Claudia Sahm
Source: Institute for New Economic Thinking
Date: MAR 2, 2021
The $1.9 trillion stimulus should be large because the need is large
The $1.9 trillion relief package is on track to pass in March but not without a struggle and with some important details still uncertain. The price tag is big, coming on the heels of the nearly $4 trillion Congress appropriated last year. That’s six times the fiscal relief in the first two years of the Great Recession. Even without the new package, the U.S. federal debt is more than GDP, according to the Congressional Budget Office, a level not seen since World War II.
With the stakes so high, disagreement among economists, even those who normally agree with each other, is heated. The question is whether spending at this level is necessary for full recovery or will instead overheat the economy. It appears that the inflation hawks have lost this skirmish, but the war is only getting started.
In the ranks of the inflation hawks are many revered macroeconomists. Most vocal are Larry Summers, a former Secretary of the Treasury, and Olivier Blanchard, a former Chief Economist at the International Monetary Fund. John Taylor, Greg Mankiw, and Bill Dudley have raised similar concerns. On the other side are Janet Yellen, current Secretary of the Treasury, Jay Powell, current Chair of the Federal Reserve, Paul Krugman, a past Nobel Prize winner, and Gita Gopinath, the current Chief Economist at the International Monetary Fund, among others.
All the sound and fury have recently unsettled bond markets. Last week, 10-year Treasury rates briefly hit 1.5% and remain at their highest levels since the pandemic began. Market-based inflation expectations moved up sharply, though their break-evens reflect other factors besides expected inflation. Nevertheless, inflation hawks point to these market signs as evidence of the risk that more deficit spending would push up rates and increase the government’s costs of servicing the federal debt. The markets steadied soon thereafter but the path forward in financial markets is far from certain. Moreover, as is always the case with markets, it is impossible to pinpoint the cause of day-to-day movements.
A critical review of the claims in this debate is plainly necessary. This essay focuses on four main areas of debate: potential output, inflation expectations, targeting on need, and pent-up demand. In each case, the data, research, and prior experience strongly suggest that the risks of doing too little far outweigh the risks of doing too much. The hawks raise valid points, but given what we know now, overheating is unlikely.
The $1.9 trillion package is not too big, it’s our estimates of capacity that are too low.
The risk of overheating rests on the dubious assumption that the relief package will create more demand than can be met easily and will thus lead to uncontrollable inflation. Recent data do not support that gloomy scenario. With ten million fewer jobs than prior to the pandemic, the case is strong to do more to support demand and a rapid recovery in the labor market.
Making the same mistake that so many did in the policy debates of the Great Recession and its painfully slow recovery, the hawks are underestimating the economy’s capacity. They base their argument on elementary comparisons of potential output from the Congressional Budget Office and GDP from the Bureau of Economic Analysis. That approach is outdated and flawed.
So, what is potential output and why is it so prominent in this debate? It’s the level of GDP beyond which inflation would pick up: the point at which demand outstrips supply and prices in general begin to rise. Once inflation picks up the fear is then that it would quickly become uncontrollable.
Conceptually, the hawks’ focus on the difference between GDP and potential output—the so-called output gap—is sensible. But they breeze past the commanding fact that similar calculations during the Great Recession understated the shortfall in demand and led to policy responses that were far too small.
Let’s look at Blanchard’s arithmetic on potential output:
“In January 2020, the unemployment rate was 3.5 percent, the lowest since 1953; it can reasonably be taken as being close to the natural rate. Put another way, output was probably very close to potential. The Congressional Budget Office (CBO) has estimated the potential real growth for the past few years at around 1.7 percent. Given that actual real GDP in 2020 Q4 was 2.5 percent below its level a year earlier, this CBO estimate implies an output gap in 2020 Q4 of 1.7% + 2.5% = 4.2%, or, in nominal terms, about $900 billion.”
His math checks out, but his assurance does not. For years after the Great Recession, estimates of potential output have been revised downward again and again. See Figure 1. For example, in its forecast published in 2010, CBO expected potential output in 2020 to be $4 trillion (in nominal dollars) higher than its estimate it published in 2020. Using CBO’s earlier forecast and the BEA’s current estimate of GDP, the output gap would be $2½ trillion. That’s more than the relief package and nearly three times Blanchard’s estimate.
The radically different views about the shortfall in output is no simple coding error. They do not reflect a rethinking of CBO’s methods over time. The dramatic differences are the direct result of years of economic distress after the Great Recession and a policy response that was profoundly lacking. It is widely understood now that the output gap was much larger in 2009 than we thought. Summers, who was then the Director of the National Economic Council, could not have known how dramatically the estimates of potential output would revise down. He could not have known how tragically small the American Recovery and Reinvestment Act would prove to be. But he and the other inflation hawks should know better now. How many times do they have to put the economic well-being of American families at risk? Yes, the time could be different. The relief package could be more than the bare minimum. But that thinking was disastrous in the past recession and is a risk too large to take again.
It is unlikely that the U.S. economy could quickly reach the levels expected in 2010. But this exercise illustrates how a timid policy response in the face of the Great Recession led to immense damage in our productive capacity. It is likely that the economy could return to its pre-Covid operation, and with ten million jobs fewer than before the crisis—a larger shortfall than any time in the Great Recession—substantial unused capacity remains.
There are other problems with focusing on current estimates of aggregate GDP. First, the headline number masks a dramatic shift in the components during the Covid crisis. Specifically, services spending, which was half of GDP in January 2021 remains $650 billion lower at an annual rate. Goods spending, in sharp contrast, is $500 higher, and the bulk of that increase comes from outlays on durables. It is striking that the rise occurred rapidly and did not result in uncontrollable goods prices inflation. The second problem is that GDP has become increasing unequal across households. In the new package, contrary to some widely publicized claims, cash relief is targeted to most families, not the wealthiest. Finally, it is important to remember that GDP is subject to revisions, which can be substantial. The BEA release states, “The GDP estimate released today is based on source data that are incomplete or subject to further revision by the source agency.” During the last recession, the downward revisions in GDP were large. Initially, the BEA estimated that GDP declined about only ¼ percent during the four quarters of 2008. The current estimate for 2008, after years of revisions, is a decline of 2¾ percent. It is unwise and at odds with past experiences to assume that the first read on GDP in 2020 fully captures the depths of the Covid crisis. Using current estimates of GDP to judge the output gap could lead us to do too little, as we did in the Great Recession.
We know from the Great Recession that revisions to actual and potential GDP can dramatically change estimates of the output gap. When the Obama Administration took office in 2009, the GDP gap was estimated to be -3%. Within a few years, as discussed in (Sahm, 2015), the estimate of the gap widened to -5%. It’s clear now that the that relief in 2009 should have been much larger. Fiscal restraint cost the American people dearly, and we must not repeat the mistake.
I am not the first to question the standard approach to output gaps. Fontanari, Palumbo, and Salvatori (2019) discuss the construction of CBO-style estimates—developed by Arthur Okun over 50 years ago (1962). They propose an alternate approach that does not depend on unobservables like the natural rate of unemployment and show that it produces better, more stable estimates of economic capacity. They argue that the massive drop in demand in the Great Recession did not lower long-run growth prospects as the CBO revisions suggest. Rather, because of the ways output is usually assessed in mainstream theory, the fall in incomes created enormous gaps relative to true potential output. They argue that the policy should have been much more aggressive in the Great Recession. Likewise, substantial relief is needed now.
Coibion, Gorodnichenko, and Ulate (2018) provide three other estimates of potential output in 2016—a time when the Federal Reserve justified the first increase in interest rates since the Great Recession on the belief that we were nearing full employment and risking overheating. Fed officials now view that step as premature. Other estimates, including Blanchard and Quah (1989), Galí (1999), and Cochrane (1994), all show markedly larger gaps and less cyclicality than those based on CBO’s approach.
Coibion et al. examine a full range of dynamic properties in potential output estimates and conclude that CBO’s framework understated how much the economy could grow in 2016 without overheating. Their analysis predates the Covid crisis by three years, but their work highlights potential problems in recent CBO estimates. To be clear, their findings do not prove that CBO’s current output gaps are too small now or then. But their research underscores the uncertainty around such estimates. Taken together, these problems suggest that it is a mistake for inflation hawks to rely so heavily on them.
Inflation expectations among families are unlikely to move up and put pressure on inflation higher.
During the past few decades, actual inflation and expected inflation among families have been low and stable. Many economists credit monetary policy for stable, low inflation and creating the expectations of a similar path going forward. See Figure 2. The stability is referred to as “anchored expectations.”