A renewed rise in Treasury yields in response to growing expectations of a large US post-election fiscal stimulus would test the Fed’s resolve to keep monetary conditions extremely loose. We suspect it would want to keep yields anchored, “talking” them down, before “walking” them down if that failed. Before considering raising the rate of its purchases, we imagine it would skew them towards longer-dated bonds.
- A renewed rise in longer-dated Treasury yields in response to growing expectations of a large US post-election fiscal stimulus would test the Fed’s resolve to keep monetary conditions extremely loose. We suspect that it would want to keep those yields firmly anchored. To achieve that objective, it could potentially either ramp the pace of its purchases and/or skew them towards longer-dated securities.
- Admittedly, pressure on the Fed to act has eased in the past week, as the 10-year Treasury yield, for example, has dropped back to near 0.7% from a four-month high of nearly 0.8%. But it is not hard to imagine another increase in the yield if expectations of a Democratic “clean sweep” build. After all, fiscal stimulus might be substantial in such an event, which does not appear fully discounted in the markets.
- We suggested here that the Fed might accept some extra rise in the 10-year yield if it were mainly due to a further rise in inflation compensation (see Chart 1), since the central bank might see less need for it to act in this instance. Yet the Fed Chair at least appears to have been calling for more fiscal stimulus and seems willing to accommodate it. So, we imagine the Fed would prefer to keep yields in general well anchored.
- One solution for the Fed, if faced with a renewed rise in long-term yields that it struggled to talk down, would be to ramp up its purchases of securities again. The Fed’s outright holdings (of Treasuries in particular) surged in the spring in response to COVID-19 but have risen more gradually since then.
- Admittedly, the Fed’s holdings have already reached an unprecedented share of economic output. (See Chart 2.) But those of the BoJ, for example, are far higher in relation to Japan’s GDP. And if the overall take-up of the Fed’s emergency lending facilities had been greater and the repayment to it of central bank liquidity swaps had not been so extensive, its balance sheet would already have been a lot larger. In fact, since early June an increase in the Fed’s outright holdings of securities has not quite compensated for a decrease in other forms of reserve bank credit, resulting in a slight shrinkage of its balance sheet.
- The Fed could also potentially skew its asset purchases towards the longer end of the yield curve, either independently of, or in conjunction with, a faster pace of asset purchases. This would be a “buy and twist” version of the “twist” operation that it conducted between September 2011 and the end of 2012.
- Back then, the Fed kept its outright holdings broadly stable rather than increasing them. (See Chart 3.) However, it significantly lengthened the average maturity of its holdings (see Charts 4 & 5), which involved selling Treasuries with maturities of less than 3 years to buy those with maturities of 6 to 30 years. The share in its portfolio of Treasuries with maturities above 5 years rose from ~47% to ~78% in this period.
- The Fed has scope to do something similar again, as this share has fallen back to ~40% now. This pull-back over time has probably reflected broadly “market-neutral” purchases by the central bank. These have been concentrated at the shorter-end of the Treasury curve partly because the outstanding value of such bonds is greater than that of longer-dated ones. For example, between 19th February (when global equities began to plummet) and 22nd April, the Fed’s holdings of Treasuries with maturities of less than 5 years rose by nearly $900bn, whereas those of longer-dated ones only rose by just over $560bn. (See Chart 4 again.)
- Even if the Fed did undertake a “buy and twist” operation, though, we doubt that it would want to bring long-term Treasury yields down substantially, since they are still very depressed in the expectation of a prolonged period of very low interest rates. Indeed, they are not a long way above our end-year forecasts. In the case of the 10-year yield, these are all 0.5% for 2020, 2021 and 2022.
- Compare the situation now with late 2011. Back then, the 10-year yield had fallen sharply before the Fed’s “twist” operation began, partly due to the brewing crisis in the euro-zone. Yet it was still nearly 1.9%. As it happens, the yield edged up in the following months after an initial drop, reaching nearly 2.4% in March the next year. (See Chart 6.) But it subsequently fell amid a re-intensification of that crisis until July 2012.
- Overall, the yields of Treasuries with maturities of 5-10 years, and more than 10 years, did generally fall slightly during the Fed’s “twist” operation in 2011-2012, while those of bonds with shorter remaining lives edged up. At the same time, though, the stock market made significant headway. (See Chart 7.) This is reflected in the relative returns from the different assets during this period. (See Chart 8.)
- We do not expect to see the US stock market outperform Treasuries as dramatically in the coming years, regardless of whether the Fed does a “buy and twist” operation. Even so, we project that the return from S&P 500 will exceed that from Treasuries in general overall between now and the end of 2022. (See here.)
Chart 1: US 10-Year Treasury & TIPS Yields
Chart 2: Federal Reserve Banks’ Holdings Of Securities As A Share Of US Nominal GDP (1914-2020*) (%)
Chart 3: Federal Reserve Banks’ Outright Holdings
Chart 4: Federal Reserve Banks’ Outright Holdings
Chart 5: Federal Reserve Banks’ Outright Holdings
Chart 6: US 10-Year Treasury & TIPS Yields
Chart 7: US Treasury Yields By Maturity Bucket
Chart 8: ~ Returns From US Treasury Yields By Maturity Bucket & S&P 500 (During Operation Twist) (%)
Sources: Refinitiv, Fed, Measuring Worth, BEA, CE
John Higgins, Chief Markets Economist, firstname.lastname@example.org