Not long ago the Federal Reserve expected to quietly shed nearly half of its $4.5-trillion portfolio by around 2022, leaving little trace of the extraordinary steps it took to face down the financial crisis. But an unexpected market kink could force the Fed to scrap the plan two or three years early and permanently leave it holding $1 trillion more than it wanted.
The U.S. central bank is making adjustments on the fly and keeping its options open. “I don’t think that’s problematic in any way” to halt the process “somewhat earlier,” William Dudley, the former New York Fed president and key architect of the portfolio strategy, told reporters last month.
Yet if the world’s largest holder of U.S. bonds tossed out its playbook and effectively took on a more accommodative stance, the result could be an across-the-board easing of market borrowing costs, the foreign-exchange value of the dollar, and of the growing strains on emerging markets.
“The evidence that we have suggests that the ultimate size of the balance sheet will be bigger than what people expected,” said Matthew Luzzetti, senior economist at Deutsche Bank Securities in New York.
All of this amounts to the final chapter in the Fed’s unprecedented decision over the last decade to buy some $3.5 trillion in mortgage and Treasury bonds in an effort to boost riskier investments, hiring and economic recovery from recession. In a nod to a stronger U.S. economy, the Fed since 2015 has raised interest rates well above zero and, since October of last year, begun shrinking its balance sheet to a more normal but yet-unspecified size.
The market kink is part of the Fed’s making.
With each dollar worth of bonds it has let run off, the so-called excess reserves it requires private banks to hold have become more scarce. Meanwhile the Trump administration’s spending boost this year has sucked up dollar liquidity as the government issued more bonds and deposited more at the central bank.
The result has been a pronounced jump in demand for excess bank reserves, which exploded during the crisis as the Fed ramped up bond-buying, but which have fallen by about $350 billion in the last nine months as it scaled back its portfolio.
That, in turn, has threatened to push the Fed’s key rate above a policy band currently set at 1.75 to 2 percent. In June, the central bank responded with a short-term fix that may only delay a longer-term reckoning.
Some economists are now predicting it will have to stop shedding bonds in one to two years and, when it’s done its post-crisis “balance sheet normalization,” end up with a roughly $3.5-trillion portfolio.
That’s larger than the broad $2.3- to $2.9-trillion range Fed economists projected in September, and well above the $900 billion it had before the 2007-2009 crisis. The September projections saw the process end by 2022 or 2023.
Sliding finish line
The looming question is how much excess reserves banks will require in the years ahead, since that represents a big yet shrinking chunk of the Fed’s liabilities. After peaking at about $2.8 trillion in 2014, they are down to $1.9 trillion.
Luzzetti at Deutsche Bank predicts excess reserves will fall to an equilibrium around $1.3 trillion, forcing the Fed to stop shedding bonds in the first half of 2020 and leaving it with some $3.7 trillion in total assets. Morgan Stanley economists on Thursday predicted it would end in September of next year.
That would spell good news for economies that have seen investment flee this year like Argentina, Philippines and India, whose central bank chief in June complained the Fed’s bond-shedding is weakening emerging-market currencies. China, locked in a trade war with the United States, could also see relief for its tumbling yuan and equities.
Closer to home, rewriting the play book could keep a lid on longer-term U.S. bond yields since the Fed would take on a more accommodative stance, perhaps permanently.
Roberto Perli, partner at Cornerstone Macro and a former Fed economist, predicted the Fed would stop shrinking its balance sheet in the second half of next year. That would mean yields on 10-year Treasuries – which fell as the Fed snapped up bonds over the last decade – would rise by only 0.25 to 0.3 percent, or about two-thirds of what would be expected were it to stick to its original plan, he said.
The so-called yield curve, or the difference between 10- and 2-year Treasury bond yields, should within about a year be completely flat – a historical signal of a pending recession. “These are not trivial amounts,” Perli said of the $800 billion to $1 trillion in additional bonds the Fed may end up holding.
To be sure, U.S. central bankers have always said their asset-shedding plan, announced in detail last June, would be flexible and depend on unpredictable Treasury issuance and redemptions of mortgage-backed securities. And the Fed has not yet specified what portion of assets banks must hold as excess reserves, which as a multiple of required reserves are near their lowest levels since the crisis.
The central bank may yet meet its forecasts, especially if the recent jump in demand for bank reserves eases.
“We don’t exactly know when we will hit that point” where excess reserves reach an equilibrium, Philadelphia Fed President Patrick Harker said at a forum in Idaho on Thursday, adding it could be between $500 billion and $700 billion.
Yet an analysis of the central bank’s annual predictions suggests it has come to terms with a permanently larger balance sheet. The New York Fed’s end-point estimates, based in part on the estimates of dealer banks, have risen to $3 trillion from $2 trillion over three years. The portfolio “normalizes” sometime between 2020 and 2022, it predicted in April.
On June 15, his last day at the New York Fed, Dudley said: “I don’t think it’s a huge problem or issue if banks demand $1 trillion of excess reserves versus $5oo billion of excess reserves, and that were to cause us to stop the balance sheet normalization process somewhat earlier.”