The Fed's Crystal Ball: Looking Beyond the COVID-19 Recession
“[W]e're not thinking about raising rates. We're not even thinking about thinking about raising rates.” (emphasis added) Federal Reserve Chair Jerome Powell, press conference transcript, June 10, 2020
Over the past 75 years, no one has seen anything like the COVID-19 shock to the global economy. Nor have we seen anything like the swift, broad and massive fiscal and monetary expansion that followed. Nevertheless, the June OECD Economic Outlook now predicts that OECD economies will contract by 7.5% this year—more than twice as large as the decline in the Great Financial Crisis (GFC) of 2007-09.
In the United States, the economic rebound has started. As states and municipalities relax the lockdown, businesses closed by the virus are gradually reopening and employment is rising. But, there remains tremendous uncertainty about the speed and extent of the recovery.
This was the backdrop for the Federal Open Market Committee’s (FOMC) release last week of its June Summary of Economic Projections (SEP)—the first SEP since December. There are two reasons why observers examine the SEP closely. First, it includes the FOMC participants’ views of the likely path of the interest rate over the coming two to three years. Second, the SEP reports projections for economic growth, unemployment and inflation, consistent with their policy paths.
Unsurprisingly, attention usually focuses on the interest rate projections, which on this occasion strongly buttressed the opening citation from Chair Powell: with the exception of two participants, the Committee does not anticipate an interest rate increase over the forecast horizon to the end of 2022.
In this post, we concentrate on the Committee’s projections for the real economy. Our conclusion is that these contain two elements of optimism. First, while the recession is clearly the worst since the 1930s, FOMC participants believe that the recovery will be roughly twice as fast as the one from the GFC. Second, their projections are that longer-run economic growth will match the pre-COVID pace. That is, in contrast to the GFC experience, COVID-19 will not usher in a slowdown in trend growth. Compared to the FOMC, we believe there is room for disappointment, especially with regard to the longer run.
We start with the policy backdrop for the June SEP. While they do not employ this exact language, the Committee is clear that they will need to continue with the current extremely expansionary policy for an extended period. In addition to keeping its key policy rate (the interest rate on excess reserves) close to zero, the Fed will sustain a variety of asset purchase and credit programs that have already increased the size of the Fed’s balance sheet by $3 trillion since end-February. (See our previous posts here, here, and here.)
A quick comparison with the 2007-09 financial crisis helps to put these actions into context. From mid-March to mid-April of this year, the Fed revived all of the lending programs they created in 2008 and 2009, and announced seven more. In addition, in the past three months, the expansion of the Fed’s balance sheet matches the entire increase from 2008 to 2013. It is difficult to overstate the importance of these actions in countering the impact of the COVID shock, keeping financial markets healthy, and ensuring that a very deep recession does not turn into a protracted depression.
On top of this massive monetary expansion is an unprecedented peacetime fiscal stimulus. Six weeks ago, the Congressional Budget Office (CBO) estimated that the combination of automatic stabilizers and discretionary spending is raising the federal budget deficit by a combined $2.8 trillion and $1.1 trillion in the 2020 and 2021 fiscal years. This corresponds to 13% and 5% of GDP, respectively. The last time such a fiscal expansion occurred, the country was mobilizing to fight WWII. (See our earlier post for a discussion of fiscal policy space.)
The economic rebound in the FOMC’s projections reflects the impact of these extraordinary policies. The following chart shows the level of real GDP (solid red), the CBO’s January 2020 estimates of potential GDP (black), and the implication of the median FOMC projections through the end of 2022 (dashed red). We plot this on a log scale, so the slope of the lines correspond to the economy’s growth rate.
Actual, potential and FOMC projections for real gross domestic product, 2000 to 2024
Using the average of the latest nowcasts from the Federal Reserve Banks of Atlanta, New York, and St. Louis, the trough of the recession in the current quarter is roughly 13% below potential. While this is more than twice as deep as the trough in the 2007-09 recession, the recovery is forecast to be much faster. Over the next 10 quarters, the FOMC projections show that the output gap will close to 4% of GDP. It took roughly twice as long―five years―to close the same fraction of the gap following the last recession. That said, even with interest rates at zero, continuing asset purchases, and a myriad of extraordinary lending programs, activity seems likely to remain well below potential beyond the 2½-year horizon of the FOMC projections.
Of course, there must be an extraordinary degree of uncertainty surrounding these projections. The FOMC does publish information on participants’ view of the distribution of possible outcomes. But, they release this material only with a three-week delay (as an annex to the meeting minutes). It is precisely at times like this when a speedier release of this information would be so valuable--something that we have been advocating for some time (see our recommendations here).
The idea that—for years to come—the Fed policy rate will be stuck at zero and that the economy will be operating with considerable slack is clearly consistent with policy simulations reported by Kiley and Roberts a few years ago. They show that, if the neutral real interest rate (r*) is 1%, then when the policy rate hits the zero bound, it stays there for 10 quarters and the output gap averages 1.1% of potential GDP. Moreover, in their simulations, conventional monetary policy is stuck at the zero bound a whopping 38% of the time.
While the Kiley-Roberts simulations exclude balance sheet policies and forward guidance, there also are reasons to think that the current episode is worse than their simulations imply. First, as noted, the COVID plunge was far sharper and deeper than anything in post-WWII experience. Second, r* is probably well below 1%: the latest FOMC projections, as well as the current estimates from the Laubach and Williams model computed by the New York Fed, put it at roughly 0.5%. That is, the shock is bigger and the conventional policy easing smaller than Kiley and Roberts consider.
So, the FOMC projections through 2022, together with Chair Powell’s press conference and the most recent Monetary Policy Report to Congress, do anticipate persistent economic slack that justifies their exceptionally expansionary posture. At the same time, the SEP outlook also looks hopeful by comparison with the recovery from the Great Recession.
Furthermore, the FOMC’s longer-run projections appear optimistic. In effect, the FOMC has interpreted all the news over the past few months as having no lasting impact on the trend rate of economic growth or on the sustainable rate of unemployment. This is visible in the following chart that plots the FOMC’s longer-run projections for growth. The red diamonds correspond to the projection medians, while the yellow and blue squares denote the top and bottom projections.
FOMC Summary of Economic Projections for the longer-run growth rate of real GDP, April 2009 to June 2020
Sources: FRED and (prior to April 2015) various FOMC transcripts. Note that the red diamond for March 2015 denotes the midpoint of the range, not the median.
As an aside, we should mention technical concerns about using the median projections both in comparing them from one year to the next and in comparing different quantities at the same time. First, the composition of the Committee changes as both Governors and Reserve Bank Presidents come and go. Thus, the medians reported from one year to the next could very well come from different people. Second, the median projection at time t of the longer run growth rate need not be the projection of the same person as the median projection at time t of the longer run unemployment rate. However, our reading of the historical evidence is that these complications are usually of secondary importance, so that analysis of the type we pursue here is useful.
Looking at the chart more closely, we note two things. Focusing on the median, from 2009 to 2012 expectations for potential growth were around 2.5%, before falling below 2% in 2016 and staying there. The latest SEP projections for potential growth continue to conform to the post-2016 pattern.
Given the post-GFC experience, is it realistic to expect that the coronavirus shocks will leave long-run growth unchanged? Following the 2007-09 recession, both the CBO and the FOMC reduced their estimates of potential GDP growth by one-half to one percentage point. Moreover, we have never returned to the pre-GFC path for the level of GDP. Given this experience, we question whether the economy will return to the pre-COVID path for the level of GDP over the next five to seven years.
Indeed, we would be surprised if the COVID shock does not further diminish longer-run U.S. growth prospects. In a recent presentation, after noting that the IMF is in the process of revising down its forecasts, Chief Economist Gita Gopinath suggests that there is a strong possibility that the depth and scope of the downturn will result in longer-term scarring. She lists three reasons: 1) the need to reallocate labor across sectors; 2) widespread firm failures; and 3) changes in consumer behavior. To these we would add the costs of making economic activity biologically safer and more resilient, and the need for reallocation of capital. Moreover, the work of Kozlowski, Veldkamp and Venkateswaran implies that changes in behavior associated with COVID will affect growth prospects for a long time.
In closing, we take two broad messages from the FOMC’s SEP and from monetary policymakers’ recent statements. First, while the recession already is very deep, the massive policy stimulus will speed the recovery compared with the post-GFC episode. Second, there will be little if any damage to the economy’s potential rate of growth.
On both points, there is room for disappointment. For example, how much can monetary policy do if the recovery were to falter? While there is much talk of “yield curve control” (see our earlier post), the SEP already is helping to anchor the two-year Treasury yield. And, with the five-year yield at 0.33% and the 10-year yield at 0.71 percent, the scope for further impact through this channel seems quite limited. Perhaps most important, however, is the possibility of long-term economic scarring. Here we turn to the fiscal authorities. So long as central banks are making credit available to healthy borrowers, it is the job of elected officials (including fiscal policymakers) to ensure that capital and labor reallocation occurs quickly, smoothly and at the lowest social cost possible.