The Fed Has No Choice But to Assume the Worst

Earlier this year, the Federal Reserve seemed to have time on its side. Payrolls were growing at a healthy clip and the unemployment rate hovered near a five-decade low. Even though there were signs that inflation was licked, there didn’t appear to be much harm in keeping interest rates elevated for a while longer — just in case. Unfortunately, policymakers can no longer take the resilience of the labor market for granted.

Perhaps the most salient detail in the August payrolls report was the net negative revision of 86,000 jobs in the previous two months’ data. The Bureau of Labor Statistics does the best job it can to deliver timely labor market data to the public, but the first drafts often end up being imperfect. Numbers are revised a couple of times as additional survey responses roll in, and there are further — sometimes larger — revisions during an annual benchmarking process. This means that the monthly data is often foggy, with large margins of error.

That brings us to last week’s news that the US added 142,000 jobs last month, which is quite fine on the face of it. All we can really say with 90% confidence though is that the actual number is somewhere between 7,600 and 276,4000, according to confidence intervals provided by the BLS. That’s a wide range! And history has shown that the revisions can be procyclical: positive in expansions, but often negative in slowdowns and recessions.

The experience of the last several years seems to echo that pattern, especially after you account for still-preliminary benchmark revisions announced last month that are expected to reduce April 2023-March 2024 payrolls gains by 818,000. All in all, a risk-management approach to policymaking demands that the Fed assume that the latest numbers are also somewhat worse than meets the eye.

downward