Peter Navarro looked at the September jobs report and called it proof that the Federal Reserve is interfering in the midterms. The report was released on October 2, 2026. The rate hike he is prosecuting happened on September 16. Sixteen days separate them. A document that did not exist cannot be the evidence a committee ignored.
That is the idea. The jobs report can change the odds of the next vote. It cannot convict the last one. Investors who hate the Federal Reserve still have to price a committee, a dot plot, and a bond yield. A matching count of days before two elections is a coincidence you can check. It is not a mechanism you can trade.
Nothing here is advice to buy or sell a bond, a stock, a fund, or a currency. A central bank can be wrong and still set the price of money. Hating the people in the room does not tell you where the funds rate will be in November.
What the report actually said
The Bureau of Labor Statistics said nonfarm payrolls rose by 29,000 in September. The unemployment rate rose a tenth, to 4.2%. Both, in the bureau's own words, "changed little." Wall Street had been looking for something near 90,000 jobs. The miss is real. "Changed little" is also real. A miss against a forecast is not the same sentence as a collapse.
The split inside the 29,000 is the part Navarro gets right, and it is in the tables. Private employers added 46,000. Government shed 17,000. The headline is a private gain minus a public loss. Over the year, federal employment is down sharply from its peak in the fall of 2024. You can call a smaller public payroll a policy success or a local recession, depending on the town. You cannot call the 29,000 a single story about "the economy."
The household survey was the stronger page. One market accounting put employment in that survey up about 406,000 and the labor force up about 485,000. The labor force participation rate rose two tenths, to 61.8%. The unemployment rate can rise when more people start looking, even if employment rises too. That is arithmetic. If 485,000 enter and about 406,000 find work, the leftover adds to the jobless count, and the rate can tick up. Navarro's line, that the tenth of a point was people coming off the sidelines, fits that arithmetic. It is not a cover story. It is how the rate is built.
Prime-age workers, those 25 to 54, are the cleanest check. Their employment-population ratio rose three tenths, to 80.7%, on the St. Louis Fed's reading of the bureau's series. For prime-age men the ratio rose to 86.2%. The Joint Economic Committee put prime-age participation at 83.7%, up three tenths on the month. People in the heart of their working years were more attached to jobs, not less. A labor market that is shedding workers does not print that.
The broader unemployment rate went the other way from the headline. U-6, which counts some of the underemployed, fell a tenth, to 7.6%. A headline rate up and a broad rate down is not a crisis print. It is a mixed print. Mixed is the word the next meeting has to live with.
Breakeven is a range, not a pardon
Navarro says the old rule of thumb, that the country needed well over 100,000 jobs a month just to hold the unemployment rate steady, is obsolete. Borders are tighter. The population is older. He puts the new breakeven near 40,000 by most estimates, and says the Dallas Fed puts it near zero. On that view, 29,000 sits "in the neighborhood."
Treat the neighborhood as a dispute, not as a fact he settled. A separate market note put the three-month average of payrolls at a revised 51,000 and called that roughly Goldman's breakeven. September, at 29,000, is under that average. It is not a boom. It is also not, by itself, a recession signal, if the bar for "enough" has fallen because fewer people are entering the country to work. Both claims can be true. What an investor needs is the range. If breakeven is near zero, 29,000 is fine. If breakeven is near 50,000, September was light. The Fed's staff has a view. The market has already traded one. You do not have to adopt Navarro's end of the range to see that the old 100,000 habit is a stale lens.
The bureau also said the prior twelve months had averaged about 45,000. September was below that average. "Well within breakeven" is a judgment. "Below the recent average, in a slowdown that was already slow" is the plainer sentence.
Where the jobs were, and where they were not
Manufacturing did add 9,000 jobs. The bureau said manufacturing employment was little changed on the month and is up 72,000 since a recent low in December 2025. Navarro calls that this year's gain and sets it against losses of more than 200,000 in the last two Biden years. The 9,000 and the 72,000 since December check out. The 200,000 comparison is his framing, and it is a different window. Use the window you can see.
Do not let the factory line become the whole report. The largest sector gain in September was not manufacturing. Private education and health services added about 20,000. Trade, transportation, and utilities added about 18,000. Construction added about 11,000, a seventh straight monthly gain in one account. Information fell. Hiring breadth was thin. A story called "the industrial turn" is a real slice. It is not the median industry. Factories matter for the next two years of production. They were not what carried September.
Navarro goes further. He says specialty trades that build factories added 12,300 in the month and nearly 112,000 since January 2025, and that real private fixed investment is running at a 6.9% annual rate. Those figures are in his piece. They are not in the employment summary this article is using. A payroll table cannot confirm an investment rate. If you want the investment claim, read the GDP tables, not a jobs op-ed. The direction he is pointing at, construction employment up alongside a small factory gain, is in the bureau's numbers. The victory lap is larger than the month.
Wages do not show a spiral. They also do not show a feast.
Average hourly earnings for all private workers rose 3.0% over the year, to $37.81. They rose a nickel on the month. Production and nonsupervisory workers rose 3.3% over the year on the hourly measure, to $32.60. Navarro talks about weekly pay instead: manufacturing weekly earnings up 5%, production workers near 6%, construction up 4.7%. Weekly pay includes hours. Hourly pay does not. A longer week can lift a weekly number while the hourly number stays tame. Both can be printed in the same month. They answer different questions.
He sets those weekly gains against a headline CPI of 3.4% and a core rate of 2.4%, and says real gains of roughly 1% to 2.5% are comfortable, with no wage-price spiral. The core figure is his citation. The hourly figure the bureau emphasizes, 3.0%, does not clear a 3.4% headline CPI. On that comparison the typical hourly raise lost a little ground to headline inflation. Against a lower core rate it would look better. Real wages in this report depend on which wage and which price you pick. What you cannot get, from either pick, is a spiral. Pay is not accelerating in a way that forces a central bank's hand by itself. That is the fact that matters for October. It is a smaller fact than "comfortably positive, solid, nothing to see."
Initial claims support the "employers are not dumping people" half of his case. New claims were 197,000 in the week ended September 26, down 1,000, and had been under 200,000 for three weeks. Reuters described that as near a 57-year low, close to levels last seen in 1969. Navarro says that, measured against the size of the workforce, claims are the lowest since the series began in 1967. The level is extremely low. The ratio-to-workforce superlative is his extra step. Low claims mean layoffs are scarce. They do not mean hiring is strong. September's payrolls already showed the difference. Firms are holding workers. They are not adding many.
The hike he is angry about had not seen this report
On September 16 the Federal Open Market Committee raised the funds rate by a quarter point, to a range of 3.75% to 4%. It was the first increase since July 2023. All 12 voters said yes. Kevin Warsh, the chair, said inflation was too high and had been too high for too long. He called the move the removal of a dose of accommodation. The statement dropped language that had blamed the inflation on supply shocks, including energy. New projections showed a large majority expecting at least one more quarter-point increase by the end of the year. The New York Times noted the decision came less than two months before the midterms, and that it put Warsh at odds with President Trump, who had wanted lower borrowing costs.
Hold that next to Navarro's charge. He says the September hike broke a rule he names for Greenspan, Bernanke, and himself: do not attack the first round of an energy spike. Wait to see if wages and other prices follow. He says this jobs report is the second-round evidence, and it shows no demand-side inflation to cure. There is a real argument inside that rule. Central banks have tightened into oil shocks before and then regretted the damage. The September statement shows the committee considered that argument and rejected it. They took the supply-shock sentence out. They decided prices were broad enough to answer with a higher rate.
You can think that was a mistake. Many readers of this site will. A mistake is not what Navarro wrote. He wrote interference. The timeline does not carry that word. The jobs report he is using as Exhibit A was published on October 2. The vote was September 16. The committee hiked on the data it had, which included a stronger August payroll print that was later revised down, and on inflation it judged too high. Friday's report can say the labor market is cooler than the forecasts of that week. It cannot say the voters looked at a 29,000 print and hiked anyway. They had not seen it.
Revisions make the point sharper, not weaker. August's first print was revised down, in one account from 162,000 to 133,000. If the committee leaned on a hotter August, it leaned on a number that did not survive. That is a reason to distrust the Fed's confidence. It is still not a reason to pretend October's tables were in the room on September 16.
The 48 days are real. The plot is not a position.
Navarro's calendar trick is accurate as a count. The midterm election is Tuesday, November 3, 2026. September 16 is 48 days before that. On September 18, 2024, the Fed cut the funds rate by half a point, from 5.25%–5.50% to 4.75%–5.00%. That was 48 days before Election Day on November 5, 2024, and it was the first cut since March 2020. Two meetings. Two elections. The same gap. He asks how, other than politics, you explain a hike on the eve of an election that the data do not support.
Here is how, without giving the institution a pass. Election Day in the United States is the Tuesday after the first Monday in November. Fed meetings sit on a published calendar that does not move for campaigns. A September meeting will often land about seven weeks before a November election. Forty-eight days is what you get when a meeting falls on the 16th and the election falls on the 3rd. You would get a similar gap in any year with that pair of dates. The rhyme is neat. Neat is not intent.
Intent would need a mechanism. A chair appointed by the president, hiking against that president's public wish for cheaper money, in a 12-to-0 vote, is a strange way to run a plot against the president's party. If Warsh were appeasing "partisan anti-Trump governors," the vote would not need to be unanimous, and it would not need to include him. Everyone voted yes. A unanimous hike is evidence of a shared inflation view, or of a shared unwillingness to dissent. It is poor evidence of a chair captured by the other faction. It is also poor evidence that the chair is a hero. He did the thing the White House did not want, and he did it with the whole committee. Motive is not on the statement. The rate is.
The 2024 cut can be criticized on its own facts. It was larger than many forecasts. It came late in a campaign. Criticizing it does not turn a date rhyme into a second crime. Investors who trade the rhyme will be early or late for a reason that will not be in the minutes. The minutes will talk about inflation. Warsh already did, in the press conference: too high, for too long.
What the report did change
The report changes October. It does not rewrite September. After the weak payroll print, traders marked down the chance of another hike at this month's meeting. A Reuters tally on October 5 put the odds of an October increase near 18%, down from about 71% a week earlier. That is the market doing what Navarro wants the committee to do: look at this report and back away from a second hike, at least for this meeting.
The dots from September still matter. A large majority of officials, at that meeting, thought at least one more quarter point would be appropriate by December. The market has walked October back. It has not been handed a new dot plot. A committee that hiked unanimously and penciled in another move can still go in December if inflation does not cool. It can also hold if the labor market keeps printing numbers like September and the next inflation print cooperates. The investor's exposure is to that fork. It is not to Navarro's verdict on who is "really running the Fed."
There is a version of his second-round rule that survives the timeline. The hike was a bet that an energy shock was becoming a general price problem. The jobs report is one test of that bet, and it failed to show a wage spiral. The next CPI is the other test. If headline inflation is about 3.4% and hourly wages are rising at 3.0%, the labor market is not the engine. If the next inflation print is still hot because of oil, the committee will have to say again whether it is fighting a shock or a trend. That is a policy argument. It is available to anyone who thinks this Fed is arrogant. It does not require a plot.
What to do with the dislike
Distrust of central banks is not a hallucination. They set a price that reprices houses, stocks, the dollar, and the interest bill on the national debt, and they do it with a lag, in a room that does not stand for election. They revise the data they used. They drop a sentence about supply shocks when the sentence becomes inconvenient. They meet on a calendar that collides with campaigns and then act surprised when people notice. All of that is enough to refuse them the benefit of the doubt.
It is not enough to skip the tables. A reader who hates the institution and then swallows a 48-day theory whole has traded one authority for another. Navarro is a political economist with a brief. He is right that the household survey was stronger than the headline, that private employment carried the month, that manufacturing is no longer the job-loss story of the prior administration's last stretch, and that weekly claims are very low. He is wrong, or at least unproven, where he turns a date into a motive. He is early where he uses October 2 to indict September 16. He is loose where weekly factory pay stands in for the hourly number the bureau puts in the first paragraph.
The useful dislike is specific. This committee hiked before it saw a soft payroll month, after telling itself inflation was no longer just energy, and it did so unanimously while the White House wanted the opposite. If the next inflation reports show the energy shock fading and wages staying near 3%, the September hike will look like the Fed doing what its critics say it always does: fighting the last price, late, and calling it principle. That outcome hurts borrowers and helps anyone who needed a weaker economy to "win" an inflation fight. It does not require the governors to be election thieves. Thieves are a story. The funds rate is a price.
What an investor can actually separate
Separate the establishment survey from the household survey. Twenty-nine thousand payroll jobs, with private up 46,000 and government down 17,000, is a slow hiring month. A labor force up by hundreds of thousands, participation at 61.8%, and prime-age employment at 80.7% is not a wave of layoffs. Claims at 197,000 agree with the second reading. Firms are not firing. They are not hiring much. Portfolios that need a recession to make their bond bet work do not have it in this report. Portfolios that need a reacceleration to make their stock bet work do not have that either.
Separate September's vote from October's odds. The hike to 3.75%–4% is done. It was 12 to 0. The dots leaned toward another move by year-end. The jobs report then cut the market's October odds from about 71% to about 18%. If you are positioned for a second hike this month, this report is against you. If you are positioned for the committee to have been political in September, this report is the wrong exhibit. The right exhibit for a political charge would be a vote that contradicted the data in hand. Make that case from August's data and from inflation, if you can. Do not make it from a table published in October.
Separate a wage spiral from a real-wage argument. Hourly earnings up 3.0% do not force another hike by themselves, not against a committee that says it is watching inflation, and not when the broad hourly number fails to beat a 3.4% CPI. They also do not prove workers are gaining. Pick the series and say so. Manufacturing weekly pay, if Navarro's 5% holds, is a sector story. It is not the national hourly story.
Separate a weaker payroll print from a weaker dollar story, or a stronger one. A Fed that stops hiking is usually a softer dollar and a friend to anything priced off easier money, including gold and longer bonds. A Fed that still has a dot-plot lean toward one more hike, and an inflation rate above 3%, is not a Fed that has surrendered. The October odds moved. The year is not over. Size the bet to the meeting you mean. "They are political" does not tell you the duration of your Treasury.
Separate France, oil, and payrolls if you hold more than one asset. The same week the jobs report was being read as a gift to doves, the euro was falling on a French budget scare and energy was still in the inflation mix that made Warsh hike. A single narrative called "the Fed is easing because jobs were weak" will be wrong on the days oil or a foreign bond spread moves the dollar. The jobs report is one input. It is the input that knocked October. It is not the only input the committee will read.
What would change the idea
The idea changes if the minutes or a dissent show that voters discussed the election calendar as a reason to hike. They have not. Warsh talked about inflation. The vote was unanimous. If a later release shows the politics in the room, the 48-day count stops being a rhyme and becomes a document. Until then it is a rhyme.
The idea changes if October's meeting hikes anyway, into this payroll print, with hourly wages still near 3% and claims still under 200,000. Then the charge that the committee is not looking at the labor data gets a real exhibit, because this time the exhibit would exist before the vote. That hike would also be the one the market, as of October 5, was only pricing at about 18%. A surprise second hike would be a policy fact. You could still argue about motive. You would not have to. The rate would be enough to reprice the book.
The idea changes if the next inflation print re-accelerates outside energy and the wage numbers follow. Then the second-round test Navarro wants goes the other way, and the September hike looks early rather than invented. Early is still a criticism. It is the ordinary one. Central banks are often early, and often late, and rarely confess which until the recession or the inflation print arrives.
The idea does not change because two Septembers sat 48 days before two elections. The American calendar did that. A trader who needs a villain more than a funds rate will keep the rhyme and miss the fork between a hold in October and a hike in December. The fork is where the money is.
The close
Payrolls rose 29,000. Private employment rose 46,000. Government employment fell 17,000. Unemployment rose to 4.2% while participation rose to 61.8% and the prime-age employment ratio rose to 80.7%. U-6 fell to 7.6%. Hourly wages rose 3.0% over the year. Claims sat at 197,000. Manufacturing added 9,000 and is up 72,000 since December 2025. That is a slow, tight, mixed labor market. It is not a breakdown. It is not a boom.
The Federal Reserve hiked on September 16, by a quarter point, to 3.75%–4%, with no dissents. The chair said inflation had been too high for too long. The committee stopped blaming supply shocks alone. Most officials thought another hike was still likely this year. The jobs report came on October 2. After it, the market's odds of an October hike fell from about 71% to about 18%. The report is an argument about that meeting. It is not a confession about the last one.
Navarro is right that the headline was a bad way to read the month, and right that a wage spiral is not in the hourly data. He is not right that a matching 48-day gap proves the hike was an election operation. The chair Trump appointed voted to make money more expensive, against the preference of that White House, and twelve people agreed. You can dislike every one of them. You can think the hike will be a mistake if oil fades and jobs stay soft. Price that mistake as a path for the funds rate. Do not price it as a crime scene. The crime scene was missing its main document. The document was still at the printer.
A note on sources and limits
Payrolls of 29,000, the unemployment rate of 4.2%, participation of 61.8%, private payrolls of 46,000, government payrolls of minus 17,000, manufacturing of plus 9,000 and plus 72,000 since December 2025, hourly earnings of $37.81 and 3.0% over the year, and production-worker hourly earnings of $32.60 are from the Bureau of Labor Statistics employment situation for September 2026 and the Joint Economic Committee's tables on that release. U-6 at 7.6% is from those tables. Prime-age employment-population at 80.7%, up from 80.4%, is the St. Louis Fed's FRED series LNS12300060. Prime-age male employment-population at 86.2% is the bureau's table A-8a. The household-survey gains of about 406,000 employed and 485,000 in the labor force, the three-month payroll average near 51,000, construction's seventh gain, and hiring breadth are from market commentary on the release, not from the bureau's lead paragraph. Claims of 197,000 for the week ended September 26 are from Reuters on October 1. The September 16 hike, the 3.75%–4% range, the unanimous vote, Warsh's inflation remarks, the dropped supply-shock language, and the dot plot are from the New York Times and Reuters accounts of that meeting. The 48-day counts are calendar arithmetic from September 16, 2026 to November 3, 2026, and from September 18, 2024 to November 5, 2024. The October hike odds near 18%, down from about 71%, are from Reuters on October 5. Navarro's breakeven figures, his weekly earnings percentages, his core CPI of 2.4%, his investment-rate claim, and his specialty-trade totals are his, from the RealClearMarkets piece republished on October 5, and are not all re-derived here. Headline CPI of 3.4% is the figure those accounts use. This page sets no target and recommends no security.