While in recent weeks there has been a material increase in Fed balance sheet normalization chatter, according to a new report from Deutsche Bank analysts, it may all be for nothing for one simple reason: should the US encounter a recession in the next several years, the most likely reaction by the Fed would be another $1 trillion in QE, delaying indefinitely any expectations for a return to a "normal" balance sheet.
As a reminder, as of this month, the duration of the latest expansionary cycle - as defined by the NBER - has reached 93 months, surpassing the 92 months of the 1982-1990 cycle, and is now the third longest in history. Should the cycle persist for another 27 months, or just under two and a half years, it would be the longest period of "economic growth" in history.
That observation has prompted speculation that at some point in the next several years, the US economy will finally succumb to what historically has been a contraction in output, aka a recession. The question, then, is how would the Fed responds. The answer, according to a Deutsche Bank analysis of the future shape of the Fed's balance sheet, is with another a $1 trillion liquidity injection, which would stop dead in its tracks any plans for a Fed balance sheet renormalization.
As the DB's Matthew Luzzetti writes in his analysis, to this point, the economic backdrop considered is relatively optimistic: the economy is assumed to not enter a recession over the next eight years and the Fed is able to normalize the fed funds rate in line with their expectations. However, given that the current expansion is already well advanced, it is possible that the economy enters a recession before the Fed's balance sheet normalizes.
Prospects for normalizing the balance sheet would be dramatically altered by a recession, potentially even a mild one. This is because with the fed funds rate still relatively close to zero, and the neutral fed funds rate potentially remaining low over the coming years, the Fed may not be able to provide enough accommodation by only cutting its policy rate. Just as in the aftermath of the financial crisis, the Fed may have to turn to another round of QE to support the economy during the next recession.
How would a recession affect the Fed's ability to normalize its balance sheet? DB explains:
We use recent analysis by the Fed Board staff to calibrate the Fed's reaction function to a recession in the next several years. This work considered how the Fed would respond to a severe recession, similar in magnitude to the financial crisis with the unemployment gap rising 5 percentage points, and with the fed funds rate already at 3%. Using the Fed's model of the US economy, FRB/US, this work concluded that the fed funds rate should be cut significantly below zero to provide enough accommodation to the economy in response to the recession. With the actual fed funds rate constrained by the zero lower bound, the paper finds that the Fed could provide similar accommodation by a combination of another QE program and strong forward guidance about the future path of the fed funds rate. The size of the required QE package is significant: $2tn.
In some ways this scenario seems too pessimistic. The magnitude of the downturn considered is likely to be more severe than the next recession for several reasons: (1) they consider a recession in line with the financial crisis, (2) there is limited evidence of substantial imbalances or overheating in the economy which should limit the depth of the next recession, and (3) monetary policy is still accommodative. Instead, a milder recession appears more likely. Conversely, it may be optimistic to assume that the fed funds rate is able to reach 3% before the next recession occurs, suggesting that the Fed could have less scope to ease monetary policy by cutting rates and therefore may have to resort to a larger QE program.
We consider an alternative scenario in which the economy enters a more mild recession in 2020 as the Fed has raised the fed funds rate to just above 3%. We assume that the recession causes the unemployment rate to rise 2.5 percentage points.
Using the results from the Fed staff's work, and assuming that there results can be extended linearly, a recession that is half as severe would require the Fed to undertake a $1tn QE program in 2020. We assume that this occurs through Treasury security purchases, with the maturity distribution of these purchases consistent with QE3. Finally, in this scenario the Fed reinvests the proceeds from its maturing MBS securities in 2020 and 2021, but does not increase its holdings.
Securities are allowed to naturally roll off beginning in 2022. In this scenario, the size of the Fed's portfolio rises above $4tn, near its record high, and it does not normalize by the end of 2024, which is the end of our horizon (Figures 7 and 8). Excess reserves remain elevated at around $800bn at that time, suggesting that it could take another three to four years – or about a decade from now – for the Fed's balance sheet to normalize under this scenario.
This is just one example of countless alternative scenarios that should be considered. The timing of normalization would be extended further into the future with larger QE programs, which would be necessary if the recession is deeper or if the fed funds rate at the start of the recession is not as high as we assume. Conversely, the required size of the QE program to combat the downturn will be smaller if the starting point for the fed funds rate is higher (taking the level of the neutral fed funds rate as given) or if the recession is shallower.
In this way, stopping reinvestment can actually increase the required size of the future QE program by limiting how high the Fed is able to raise the fed funds rate prior to the recession. This, in turn, raises the likelihood and size of an eventual QE program in response to the next recession. The magnitude of this effect is likely to be substantial in the Fed's view, given that they place considerable emphasis on how it is the stock of their assets that eases financial conditions by compressing term premia. In a footnote to a recent speech by Chair Yellen, it was noted that the declining maturity of the Fed's portfolio in 2017 would raise 10-year Treasury yields by about 15bp – an effect equivalent to about a 50bp increase in the fed funds rate. Thus, the trade-off between tightening via the fed funds rate or through reducing the balance sheet is a non-trivial consideration.
These results suggest that to normalize the size and composition of its balance sheet over the longer-run, the Fed's most important objective may be to help the economy avoid a recession. Tightening financial conditions due to the start of unwinding the balance sheet may limit the extent to which the fed funds rate can increase prior the next recession, increasing the likelihood and size of the next QE program and thus delaying normalization of the balance sheet.