If nothing else, quantitative easing has constituted a full-employment act for monetary economists. More than a decade on from the birth of the QE era we are still debating what effect it had, if any.
A quick search on SSRN yields over 1,000 papers on QE, and over 500 on the National Bureau of Economic Research’s website. Then there are the millions of sellside research reports, think-tank papers, comment pieces and hedge fund letters (and, cough, a few AV posts and comments).
Conclusions range from QE being the economic equivalent of crack cocaine that at best only creates “a financial fantasyland” and “underwrites inequality”, to it saving the world from financial cataclysms and even making us bonk more.
Strong early contender for paper of the new decade. Easier monetary policy led to more post-crisis babies, Bank of England estimates. https://t.co/AyaLTtMDtJ Your best liquidity and stimulus jokes in the comments please.
— Robin Wigglesworth (@RobinWigg) January 2, 2020
The size and perhaps especially the duration of the post-Covid stimulus is particularly controversial, given that stimulus this time actually did seem to lead to faster inflation. (FTAV suspects this is almost all because of fiscal policy and global supply chain issues then compounded by a huge systemic energy shock, but anyways.)
A new paper by Andrew Levin, professor of economics at Dartmouth and visiting scholar at the IMF, his undergraduate student Brian Lu and the Bank Policy Institute’s chief economist William Nelson has explored the costs and benefits of this “QE4” programme. They are unimpressed:
QE4 was initially aimed at mitigating strains in markets for Treasuries and agency mortgage-backed securities but was subsequently aimed more broadly at supporting market functioning and providing monetary stimulus. Nonetheless, QE4 did not have any notable benefits in reducing term premiums. Moreover, since the securities purchases were financed by expanding the Fed’s short-term liabilities, QE4 amplified the interest rate risk associated with the publicly-held debt of the consolidated federal government. Our simulation analysis indicates that QE4 is likely to reduce the Federal Reserve’s remittances to the U.S. Treasury by about $760 billion over the next ten years.
Let’s unpick this a little. That the truly massive dose of stimulus the Fed unleashed when the pandemic stuck — it bought almost $2tn of bonds between March and June 2020 — undoubtedly helped avert what could have been a ruinous financial crisis on top of twin health and economic crises.
Levin, Lu and Nelson concede this, and therefore explicitly treat the initial salvo as distinct from later bond purchases that were mostly to balm the economic pain of lockdowns. But they argue that the impact was negligible and the longer-term financial losses that will accrue will hurt taxpayers. Here are their main findings:
— Program Design: The evolution of the QE4 program was opaque and inertial. Moreover, the FOMC minutes did not report any substantive discussions of cost-benefit analysis at any stage of the program, as though the costs were minor and the benefits were clear-cut.
— Consequences for Market Functioning: The Federal Reserve’ actions at the onset of the pandemic helped stabilize markets for Treasuries and MBS. Over time, however, QE4 continued to expand the Federal Reserve’s outsized footprint in those markets, which could substantially reduce market liquidity going forward. Indeed, the SOMA now holds nearly 30% of the outstanding stock of Treasury notes and bonds and more than 40% of the total outstanding stock of agency MBS, and its QE4 purchases comprised nearly the entire issuance of agency MBS over the period that the program was being conducted.
— Balance Sheet Normalization. Our baseline projection indicates that the size of the Federal Reserve’s balance sheet will reach a trough in late 2024 and then resume expanding to meet policymakers’ criterion of providing an “ample” supply of reserve balances. However, the composition of the SOMA’s asset holdings will remain far from normal, with a small proportion of Treasury bills and a glacial pace of agency MBS runoff.
— Interest Rate Risk. By purchasing medium- and longer-term Treasuries and financing those purchases by creating short-term interest-bearing liabilities, the FOMC incurred substantial interest rate risk, i.e., risk to the net interest income of its balance sheet. The FOMC’s purchases of agency MBS were associated with even greater risk because mortgage prepayments decline sharply in response to increased mortgage rates.
— Implications for Consolidated Federal Debt. The FOMC’s actions substantially reduced the average maturity of the interest-bearing liabilities of the consolidated federal government sector (which includes the Federal Reserve). Thus, while the U.S. Treasury was issuing notes and bonds to “lock in” low interest rates and reduce the expense of financing the federal debt over coming years, QE4 practically canceled out those efforts.
— Cost to Taxpayers. Based on the term structure of interest rates at the end of June 2022, our baseline projection indicates that over the next ten years the Federal Reserve’s total net interest income and its corresponding remittances to the U.S. Treasury (and hence the federal government’s total net revenue on a consolidated basis) will be about $760 billion lower than in the counterfactual scenario with no QE4 purchases. Moreover, only a small portion of that cost (about $120 billion) is associated with securities purchases when the Federal Reserve was serving as market-maker of last resort at the onset of the pandemic.
— Assessment of Benefits. The QE4 program did not have any significant effect in reducing term premiums and hence does not appear to have contributed to the very rapid pace of economic recovery in 2020-21.
Some of this looks a bit . . . unpersuasive? Here are some initial thoughts on their criticisms.
The idea that the Fed should have waited to conduct a detailed and transparent cost-benefit analysis when first rolling out the stimulus in March 2020 seems ludicrous, for example. Speed and scale were of the essence.
Just because the FOMC meeting minutes don’t feature detailed subsequent discussion as the stimulus was extended doesn’t mean that it was never discussed by the board or staff either. And this is probably the most discussed and dissected issue of monetary economic of the past decade. What more was there to say? The Fed thinks it works, ergo they did it.
The controversial “cost to taxpayers” is also bit of a mirage, as we’ve written before. Firstly, the Fed has already sent Treasury $869bn of profits from earlier QE programmes. You can’t just look at the L part of the P&L. Secondly, who really cares anyway? Normal accounting rules don’t apply to central banks. The Fed can create money and operate with negative equity. It’s not a hedge fund. It sets policy to modulate the economy, not to turn a profit.
The idea that the average maturity of the consolidated US public sector debt has been shortened also seems inconsequential. There is no rollover risk! Treasury can always lengthen out maturities further again if it wishes, but doesn’t actually try to time lows in yields and “lock in” low interest rates anyway. Otherwise there would have been helluva lot of expensive 30-year Treasuries issued in 2009, and in 2010, and in 2011, and in 2012 etc etc . . .
Lastly, solely looking at term premiums as the only gauge of any economic impact also seems a bit simplistic. One can certainly have an argument about whether the Fed should have curtailed its purchases much sooner, when growth rebounded strongly in 2021, inflation was clearly firming up and becoming problematic. But there are myriad direct and indirect ways that QE4 likely contributed to the vim of the economic recovery.
Anyway, take a look at the full paper and let us know your own thoughts.