The Federal Reserve’s Open Markets Committee, or FOMC, meets September 15 and 16 to decide whether to raise interest rates, lower interest rates, or leave them the same. I suspect Fed Chair Kevin Warsh will leave them the same. If he does so, the business press will portray that as a craven surrender to President Donald Trump. But if you think, as I do, that the August jobs report was a fluke, then leaving rates where they are is a good call.
The two factors likeliest to drive the Fed decision are job creation and inflation. Let’s start with jobs. As I pointed out on the eve of Labor Day, the labor market has been weak. During the 12 months preceding July 2026, the Bureau of Labor Statistics, or BLS, reported that the average number of jobs created was 31,000. BLS logged 63,000 jobs created in May, 31,000 in June, and 21,000 in July. Then, in August, BLS reported that job creation suddenly jumped to 162,000. If the previous slowdown in job growth really reversed itself that dramatically in August, then yes, that’s a strong argument for Warsh to raise interest rates. But did it?
Explanations offered thus far for why the dwindling-jobs trend abruptly transformed into a robust-job-creation trend have not been persuasive. For example, The New York Times’s Sydney Ember said the August job gains “come at a time of broader economic resilience, marked by strong corporate profits and vigorous capital investments spurred by A.I.” But corporate profits and capital investments in AI were similarly strong in May, June, and July, when job growth was declining. Why didn’t job creation take off then? The biggest August jobs increase was the 62,000 jobs created in the leisure and hospitality industry, which lost 21,000 jobs in July. Did everybody put off their vacations this year until August? Not according to the Leisure and Travel Association, which says travel spending and room demand were fairly robust in July.
I’m not suggesting any sort of conspiracy. I don’t think the Trump administration tampered with the jobs statistics; that’s actually quite difficult to do. But Trump has taken a meat cleaver to all federal statistical agencies, making it much harder for them to their jobs well. The Bureau of Labor Statistics, which compiles the jobs numbers, has about 20 percent fewer staff since Trump’s second term began, according to a recent report by the American Statistical Association. That’s bound to affect performance.
“BLS,” the American Statistical Association report said, “has not had the funding increases necessary to modernize its core surveys that measure critical indicators on jobs, unemployment, prices, and expenditures.” As a consequence, BLS surveys are becoming less reliable. Notably, response rates to the BLS’s monthly establishment survey are falling. This is the survey that supplies the agency’s job-growth numbers. One result is larger after-the-fact revisions than in the past.
When there’s a big one-month swing in the jobs numbers, as there was in August, it isn’t crazy to wonder whether some big adjustment in the other direction lies ahead. Commenting on the August jobs report on CNBC, Wendy Edelberg, former chief economist at the Congressional Budget Office, said: “I think that these strong payroll numbers are absolutely going to get revised down. I think we don’t have the population growth to support this kind of payroll growth.” A more likely job-creation number for August, Edelberg said, was 20,000 jobs, and “maybe zero.”
Let’s move on to inflation. The Consumer Price Index, or CPI, is another statistic compiled by the beleaguered BLS. Last year the agency cut back on its data collection for calculating CPI. “When the government can’t collect data on prices,” Ben Casselman explained in The New York Times, “it has to fill in the gaps with a statistical technique called ‘imputation.’ The more data that must be imputed, the less reliable the overall numbers become.”
The August BLS inflation report showed the Consumer Price Index climbed to 3.4 percent in August. We have less reason to doubt that than we do BLS’s August jobs nuimber, because we know that the Iran war has been driving up the price of oil. You probably won’t be surprised to learn that in August the price of gasoline was up 27 percent over the previous year. But “core inflation” (i.e., inflation minus food and energy) rose a manageable 2.4 percent over the previous 12 months. Democratic politicians should still shout from the rooftops that the CPI is higher now than it was when Trump took office (2.8 percent CPI), and that the price of gasoline actually fell during Biden’s last year in office. Trump richly deserves such abuse because during the 2024 election Trump went around saying, untruthfully, that Biden unleashed “the highest inflation perhaps in the history of our country.”
But core inflation is lower today than when Trump entered office, and the Fed tends to make its decisions based on core inflation. As it happens, the Fed’s preferred metric is not the CPI at all but the Commerce Department’s Personal Consumption Expenditures (PCE) Price Index, which shows both broad-based inflation and core inflation mostly falling lately (though the latest calculation goes only through July; August may tell a different story). The PCE percentages (3.7 percent for all items, 3.3 percent for core inflation) are higher than the CPI percentages, but if these didn’t prompt the Fed to raise interest rates during the past year (it has not, in fact, raised rates since July 2023), then it’s hard to see why the Fed should raise rates this week.
The best argument for a rate hike is that it might cool the bond markets, which pushed up 10-year Treasury bills to 5 percent on Monday. But would that really do the trick? Although the bond market is worried about inflation, it’s much more worried about the budget deficit, about which Trump can be counted on to do fuck-all.
If you think I’m letting Warsh off easy, I am not. He will still get crap from Trump for not lowering interest rates, which is what that madman wants. Also, this is not the last Fed decision on interest rates prior to the midterms. The FOMC will convene again on October 27 and 28. If my analysis here proves wrong—if the next jobs report, on October 2, confirms a two-month trend of 100,000-plus jobs created, and if the next PCE report, on September 30, shows core inflation rising significantly—then yes, Warsh will probably be well advised to raise interest rates. It will be the worst possible time so far as Republican congressional candidates are concerned. If Warsh did raise interest rates next month, Trump might try to fire Warsh and replace him with his economic yes-man, Kevin Hassett, director of the National Economic Council. But the lunacy of such a spectacle would boost Democrats in the midterms. Almost everything Trump does seems to have that effect.
