The FOMC Has Strong Reason to be More Worried About Jobs than Markets Reflect

• Nonfarm payrolls were mildly higher…

• ...but weaker under the hood

• US private payrolls ex-health care have been down all year

• The FOMC should be more concerned about the state of the job market

• Scotiabank Economics ranked #1 among 71 nonfarm payroll forecasters in 2025

• US payrolls m/m 000s / UR %, December:
• Actual: 50 / 4.4
• Scotia: 25 / 4.5
• Consensus: 70 / 4.5
• Prior: 56 / 4.5 (revised from 64 / 4.6)
• Two-month revision: -76k

Nonfarm payrolls grew by 50k which split the difference between consensus and my estimate. Not bad, right? Wrong.

For starters, it’s a minor point, but revisions removed 76,000 jobs from the prior two months and mainly from October that fell 173k before gains of 56k and 50k in the next two months.

The details leave a sour taste behind. For one thing, there was narrow breadth (chart 1). Take health sector hiring out of the picture and there isn’t much else. In fact, chart 2 shows that nonfarm has been down in four of the past eight months and roughly flat in three others with only September posting a small gain. The same holds true if we just look at private payrolls minus healthcare (chart 3).

Monthly breakeven estimates for job growth are lower in the context of tighter immigration policy, but it’s a stretch to say the payroll numbers ex-healthcare should be trending negative. And yet, nonfarm ex-health since April has been down by about –354k. That’s an extremely important observation from a monetary policy standpoint: health care hiring is less likely to be rate sensitive than the rest of the labour market, although other drivers like massive non-Fed policy uncertainty and AI haven’t helped.

The FOMC should therefore view the full employment side of its dual mandate as under more pressure than the inflation side for now. That should motivate a cut on January given the Congressional mandate, even if it’s a plug-your-nose-on-moral-hazard cut by potentially emboldening other aggressive policies.

President Trump posted aggregated ytd numbers last evening (here) that showed decent private sector job growth in 2025. Tisk tisk. He gets the numbers from the White House Director of the Economic Council late the day before but is supposed to sit on them. There’s just one small problem—it’s all been in the health sector (chart 4). Take health out, and nonfarm private payrolls were flat in 2025. 

Then consider the expiration of ACA subsidies at the end of 2025 and it’s not hard to imagine downside risk to health sector hiring in 2026. Trump clearly hates many government jobs and isn’t a fan of health care subsidies which the data suggests leaves nothing much else to drive the US labour market.

The numbers are weaker yet upon consideration of additional points. One is that the payroll figures do not yet include the effects of annual benchmarking numbers. That takes two forms. One is that next month—with the January report—the annual revisions to March 2025 payrolls will be finalized and incorporated in the official nonfarm payroll readings and they were initially estimated in September to take the March level down by 911k. Second is that from April onward there is an ongoing overestimation bias as Fed Chair Powell has indicated when he said he thinks payroll changes since April have been overestimated by about 60k per month.

If that overestimation since April holds—and we won’t get the annual benchmarking revisions for the period from April 2025 to March 2026 until September while relying upon quarterly estimates in the meantime—then the figures shown in charts 2-4 are likely considerably weaker yet. Chart 5 offers an attempt at showing how much weaker by plotting official payrolls and payrolls adjusted for these points on including revisions. 

Further, seasonal adjustment factors were not quite as high as the prior three months of consecutive record SA factors when comparing like months in history, but chart 6 shows the BLS still went to the high side. Chart 7 shows that this kept the payroll change positive versus what would have happened at other SA factors. Those SA factors are calculated using standard statistical methods but the opaqueness leaves questions behind in terms of judgement around whether SA factors should be as high as they have been ever since Trump fired the BLS Commissioner.

Where to from here? Consumers may have the answer (chart 8). And we’re getting perilously close to a worse outcome for the economy before even incorporating any of the arguments above about overstated payrolls (chart 9, grey bars are recessions).

The next page offers additional charts for your consideration. The unemployment rate edged lower to 4.4% from 4.5% (revised from 4.6%) as the more volatile household survey posted a solid job gain and a mild contraction in the workforce. The noise on this survey blows nonfarm’s statistical noise out of the water. Further, a tight labour market is one thing; job growth and income growth are what drive consumption growth.