Chairman Powell's Duration Problem
May 9, 2022 | Watching the huge fuss coming from the ranks of professional equity managers, you’d think we are in some sort of financial meltdown a la 2008. In fact, stock prices have returned to pre-COVID levels, hardly a catastrophe. Considering that the FOMC has yet to actually do anything save raise the target for Fed funds 50bp, the reaction of the markets seems a tad overdone. After all, the FRBNY will still be reinvesting the redemptions and prepayments on the system open market account or SOMA through June.
As we noted in our comment last week, Fed Chairman Jerome Powell’s press conference was more interesting than usual. Powell, you see, actually understands much of the financial markets subject matter that is so pressing upon the considerations of the FOMC. Sometimes he says things in person or in the Fed transcripts that are quite important, but these little gems are usually missed by the generalist media that covers the central bank. Powell said with respect to the runoff:
“Consistent with the principles we issued in January, we intend to significantly reduce the size of our balance sheet over time in a predictable manner by allowing the principal payments from our securities holdings to roll off the balance sheet, up to monthly cap amounts. For Treasury securities, the cap will be $30 billion per month for three months and will then increase to $60 billion per month. The decline in holdings of Treasury securities under this monthly cap will include Treasury coupon securities and, to the extent that coupon securities are less than the monthly cap, Treasury bills. For agency mortgage-backed securities, the cap will be $17.5 billion per month for three months and will then increase to $35 billion per month. At the current level of mortgage rates, the actual pace of agency MBS runoff would likely be less than this monthly cap amount.”
While the FOMC is able to plan the runoff of its Treasury portfolio with precision, the variable nature of the duration of the Fed’s MBS portfolio creates a lot of uncertainty. Thus the last sentence of Powell’s statement confirms that the natural rate of runoff of the $2.7 billion MBS portfolio is well-below the $35 billion monthly cap, suggesting to many Wall Street bond analysts that outright sales will be required next year. Since much of the Fed’s portfolio is comprised of agency and government MBS with 2% and 2.5% coupons, any sales will imply a loss of 10-15 points for the FOMC’s account.
Naturally the little nuance about losses on the portfolio flew right over the heads of most journalists at the FOMC presser, who as a group prefer the vague but promising world of monetary policy to the mechanics of the financial markets. Thus when Powell told the audience that the FOMC is essentially flying blind when it comes to adjusting the three variables of policy, few journalists in the room reacted. Powell said:
“So typically in a recession, you would have unemployment. Now you have surplus demand. So there should be room in principle, to reduce that surplus demand without putting people out of work. The issue will come that we don’t have precision surgical tools. We have essentially interest rates, the balance sheet, and forward guidance, and they’re famously blunt tools. They’re not capable of surgical precision.”
Much like the Russian bombs and missiles being employed in the pacification of the Eastern Ukraine, Fed monetary policy is a very blunt tool. Especially when the FOMC is unsure how to calibrate changes in the SOMA with other policy tools such as rate targets and forward guidance. Forward guidance does not count for much when a crowd of equity investors is headed for the door all at once. If you think for a moment, once we dispense with forward guidance, the two remaining components of FOMC's proverbial toolkit hardly inspire great confidence.
During a response to a question from no less than Michael McKee of Bloomberg News, Powell let slip the proverbial bomb that zoomed high over most heads in the room. He basically told the global markets what the readers of The Institutional Risk Analyst have known for years, namely that once you go down the dark road of massive bond purchases via quantitative easing or "QE," you cannot retrace your steps without potentially horrid consequences. Powell said:
“I would just stress how uncertain the effect is of shrinking the balance sheet. You know, we, you -- we run these models, and everyone does in this field, and make estimates of what will be the -- how do you measure, you know, a certain quantum of balance sheet shrinkage compared to quantitative easing? And, you know, these are very uncertain. I really can't be any clearer. There won't be any clearer. You know, people estimate that broadly on the path we're on, and this is -- this will be taken, probably too seriously. But sort of one quarter percent, one rate increase over the course of a year at this pace. But I would just say with very wide uncertainty bands, very wide.”
As we and our friends at ZeroHedge noted some years ago, the Fed has been acutely aware of the “duration problem” caused by QE since 2016. This technical issue has to do with the variable nature of the maturity of mortgage backed securities, but also in the way in which central banks manage their securities portfolios overall. We wrote in The Institutional Risk Analyst back in 2017 (“Banks and the Fed’s Duration Trap”):
“Since the Fed and other sovereign holders of MBS do not hedge their positions against duration risk, the selling pressure that would normally push up yields on mortgage paper and longer-dated Treasury bonds has been muted. Thus the Treasury yield curve is flattening as the FOMC pushes short-term rates higher because longer-dated Treasury paper, interest rate swaps or TBA contracts are not being sold, either in terms of cash sales by the FOMC or hedging activity.”
Like the big banks post London Whale and the Volcker Rule, the Fed’s portfolio is totally passive. There is no attempt to manage for duration or hedge, much less the more obvious public policy goals of regulating inflation. This is an odd situation given that Board staff has explicitly recognized the impact of changes in the size and composition of the SOMA.
As we said to David Andolfatto et al on Twitter a while back, maybe the FOMC should manage the SOMA for a duration target. Publish same in the Minutes? We were the first writers in 1993 and the good works of Rep. Henry B. Gonzalez (D-TX). Read about those years in Tim Todd's book, "A Corollary of Accountability: A History of FOMC Policy Communication."
Rather than merely selling securities for cash, an obviously painful policy choice, why not instead think about changing the duration of the Fed’s portfolio? That swap with the Bank of Japan we suggested earlier beckons. That said, any embedded loss in the Fed’s bond portfolio should not be a major concern for analysts who truly appreciate the relationship between the US Treasury and the central bank. The fact is that QE is and always was an expense to the Treasury, proof positive that the two legally separate agencies are actually faces of a single godhead.
Robert Eisenbeis of Cumberland Advisors clarified the issue in a December 2017 interview (“The Interview: Bob Eisenbeis on Seeking Normal at the Fed”):
“The Fed almost by definition cannot make a profit. It baffles me how people inside the system can fail to see the accounting reality here. The Fed issues short term liabilities to buy Treasuries taking duration out of the market. The Treasury makes interest payments to the Fed who takes out its operating costs, including interest payments on reserves and returns the remainder to the Treasury. If this intra governmental transfer were settled on a net basis like interest rate swaps, there would always be a net payment from the Treasury to the Fed.”
Notice that Eisenbeis, who headed Research at the Atlanta Fed, properly identifies that the Fed has withdrawn duration from the market. But the thing that rightly worries Chairman Powell and other properly focused members of the FOMC is the uncertain calculus that attaches to attempts to manage down the size of the Fed’s balance sheet. Let’s review how this works, first with QE and then quantitative tightening or QT.
With QE, the Fed of New York purchases securities in the secondary market from the primary dealers, part of the illusion of separateness between the Fed and Treasury. The Fed takes delivery of the security and credits the reserve account of the bank, which in turn records a new “deposit” either for itself or a customer. The reserves at the Fed earn interest and are cash for all purposes.
As of the May FOMC meeting, the Fed has stopped buying new securities for the SOMA, but is still reinvesting redemptions of Treasury debt and MBS, including prepayments and insurance payments from the GSEs and FHA, VA, etc. In terms of cash, the mortgage related payments to the SOMA ultimately are paid by the Treasury. Importantly, so long as the Fed continues to reinvest redemptions, the Treasury does not need to refinance the bond in the private market.
With QT, the Fed stops reinvesting principal and interest payments. The Treasury redeems the bond and gives the Fed cash, which reduces its balance sheet by a like amount of reserves. The deposit on the books of the bank disappears, however, because the Treasury, which is running a deficit, must now refinance that bond in the private market. The bank or a customer buys the new Treasury bond, but the deposit disappears and the bank shrinks.
Part of the difficulty in figuring out how to manage balance sheet shrinkage is the fact that as the Fed’s balance sheet runs off, banks must start to migrate away from reserves at the Fed for liquidity purposes and back into Treasury debt and MBS. Banks and other investors hedge these positions, creating new selling pressure in longer-dated securities at just the moment when selling pressure is already soaring. Taking down the SOMA is a process that is, by definition, problematic.
Note in the chart above that the Treasury General Account at the Fed, which is another factor in the FOMC monetary policy mechanics, is back up to just shy of $1 trillion. This account is collateralized with Treasury debt. Meanwhile, do note that the short-term money markets are still signaling deflation, with the bid on the Fed’s reverse repurchase facility falling through the market volatility last week as demand rises.
"Overnight funding rates plunged below the Federal Reserve’s new target range on Thursday as cash returned, overwhelming a market that’s already short on investable assets," reported Alexandra Harris of Bloomberg News.
She reports that GCF repo rates fell to just 0.6% last week. "Demand for the Fed’s overnight reverse repo facility rose Thursday, as 86 counterparties parked $1.85 trillion, just off the all-time high of $1.91 trillion reached on April 29."
As the SOMA shrinks, the $8 trillion in bond market duration supported by the Fed will shift back to the banks and primary dealers, especially in longer dated Treasury paper and MBS. We worry that Chairman Powell and his colleagues are still focused on managing a difficult financial process with SOMA as though it were a monetary policy narrative. The two things are not the same, as we proved in September 2019 and December 2018.
Chairman Powell’s demeanor during the press conference when he spoke with Mike McKee suggests a man who knows that he has a problem with the balance sheet. If the Fed starts to actually reduce its portfolio next month and shifts this duration back into the hands of banks and private investors, the weight of duration on the system will increase selling pressure in the bond market in ways that defy expectations. Meanwhile, buying pressure in the short end of the Treasury curve says that deflation remains the ultimate issue facing the FOMC.