Inflation surprises hurt
U.S. supply-chain stress has intensified because of China’s zero-COVID tolerance policy, and this could limit, but not eliminate, the disinflation in U.S. goods prices over the next couple of months. The Federal Reserve has zero tolerance for upside inflation surprises, and recent rhetoric from the central bank suggests it is willing to do anything to break inflation.
Financial markets also don’t like inflation surprises. The timeliest metric to assess the accuracy of inflation expectations are one-year-ahead inflation swap forecast errors.
Since inflation picked up in March 2021, inflation swaps consistently have underpredicted inflation with an average forecast error 2.5 times its historical value. However, this underprediction signals faith in the Fed’s initial stance that inflation would be transitory.
After all, unlike in 1979, the Fed now has goodwill with a 40-year price-stability record. Nevertheless, as prices rose persistently in 2021, inflation swaps gradually rose above 4% when the Fed itself pivoted to tightening in November.
When the Federal Open Market Committee meeting in March predicted 2022 inflation of 4.5%, one-year inflation swaps traded at 5.5%. Higher maturities still indicate that markets expect inflation to decelerate over the next two years.
By conventional wisdom, stable monetary policy, or the absence of inflation surprises, creates market stability, which does not necessarily mean that markets will do great. Even if supply-side pressures relax and the Fed’s anticipated policy reduces inflation to around 5% year over year by year’s end, such policy has real economic costs.
Higher interest rates have already pushed up mortgage and corporate bond rates, dampening housing and investment demand. Equity prices are unlikely to rebound due to slower earnings growth.
The downside risks are larger if the Fed fails to contain inflation. We organized median quarterly real yields and yearago returns for Treasuries, corporate bonds, stocks and housing from 1970 to 2021 by quartiles of inflation forecast errors.
Historically, the typical asset class performed significantly worse during periods of positive inflation surprises. For instance, ex-post real interest rates throughout much of the 1970s were negative, eroding savings, while the S&P 500 ended the decade where it started, painting stocks as a poor investment in a stagflation environment.
The notable exception is real estate. However, because of low affordability, real estate will at least not be an accessible hedge for everyone. Containing inflation in the coming months, costly as it may be, will thus be paramount for financial markets.
Unhappy U.S. consumers spend anyway
Second-quarter U.S. GDP growth is looking better after April retail sales and industrial production. Retail sales were up 0.9% in April, a little lighter than our forecast of 1.2%, but the details were solid. Nonauto retail sales rose 0.6%, in line with our expectation.
Nonstore retail sales rose 2.1% in April after rising only 0.4% in March. The recovery in retail sales at restaurants continued, with sales rising 2% between March and April. This is consistent with the improvement in the weekly data from OpenTable. Spending at gasoline stations fell as there was some relief at the pump. Gasoline prices atypically fell in April.
Sales of the important control group—the total excluding autos, gasoline, building materials and food services— increased 1% in April and there was a positive upward revision to March. Control retail sales were up 9.1% annualized over the prior three months in April. Some of the growth is attributed to higher prices as retail sales are not adjusted for fluctuations in prices.
However, even adjusting for price effects, real control retail sales posted decent growth over the past three months. This raised our high-frequency GDP model’s tracking of real consumer spending in the second quarter to 4.6% at an annualized rate.
The solid gain in spending in April may seem inconsistent with the drop in some measures of consumer sentiment. For one, the correlation between changes in monthly consumer sentiment and real consumer spending is low.
The correlation coefficient for the University of Michigan survey is 0.27. This is less than the 0.33 correlation coefficient between changes in real consumer spending and fluctuations in the Conference Board’s consumer sentiment index.