U.S. Manufacturing PMI Expands to Highest Level Since May 2022, Signaling Industrial Recovery

The Analysis

The ISM manufacturing index is at its highest since May 2022. That sentence is carrying a lot of weight in the freight commentary, and some of what is being loaded onto it will not fit. Strictly, a diffusion index counts directions. It does not measure size.

What we have is a relative position and three components: new orders expanding at the fastest pace in forty months, supplier delivery times lengthening, and a committee chair saying the wait-and-see posture is fading as order books fill. No index level is given. That absence matters more than it looks, because the line separating expansion from contraction sits at a fixed number, and distance from a fixed line is not the same thing as distance from the last peak.

If you buy or sell physical goods across a border, the question is not whether American industry is expanding. It is which of those three components is driving the reading, because they point at different things for your inbound and your outbound flows.

The facts are compact. The ISM manufacturing PMI printed its highest reading since May 2022, described as a broad-based recovery after a long contraction. The chair of the manufacturing survey committee, Susan Spence, said companies remain cautious under trade policy uncertainty, but that the wait-and-see posture is diminishing as order books fill and production schedules firm. New orders expanded at the fastest pace in forty months. Supplier delivery times lengthened. For logistics, the read-across is more freight demand across trucking, rail, and parcel, on top of a market already tightening from supply-side capacity reductions, and the reading is offered as macro confirmation of what rail intermodal and trucking spot data have been showing.

Now the definitional premise, because everything else rests on it. This is a diffusion index built from surveyed purchasing managers reporting whether an item went up, down, or stayed the same. A reading above the dividing line means more respondents reported improvement than deterioration. It does not say by how much, and it does not say how much volume moved. The same headline can come from half the panel reporting better and half reporting worse, or from the entire panel reporting no change at all, and those two economies look nothing alike on a dock. That is why the level should not be scaled into a capacity plan. It is a direction and a breadth signal, and it is being used as a magnitude signal in most of what I have read.

The delivery time component is where the sign flips, and this is the part I would insist on. In the construction of the index, slower supplier deliveries count as expansion, on the reasoning that purchasers are waiting because demand is outpacing supply. Read that same sentence from the buying side and it says lead times are extending, allocation is starting, and expedite fees are coming. One number, two signs. A procurement team that reads lengthening delivery times as a healthy demand signal has the arithmetic backwards for its own inbound program, and that mistake is expensive in exactly the quarter when it is easiest to make.

Composition is the only part with information in it. New orders at a forty-month fastest plus delivery times lengthening is demand outrunning supply, which is the classic signature of a market about to allocate. New orders rising with delivery times shortening would be a demand recovery with slack in it, and that is a different supply chain problem altogether: you solve the first with buffer and contracts, the second with patience. The headline cannot tell the two apart. Two lines on the same chart can.

The impact lands unevenly. Large importers on contracted carriage feel it at renewal, as tender rejections and as a contract-to-spot spread that closes against them, three to six months out. Mid-size importers feel it sooner, in spot rates and in equipment availability at origin. Small shippers feel it as a market where nobody returns a quote, and then as minimum charges. Direction matters: US importers face inbound tightening, US exporters face container and railcar availability plus the box imbalance, and exporters in Asia face booking windows that shorten before the rate does. Scope matters more than most people allow. This is a manufacturing survey. If seventy percent of your book is apparel, perishables, or parcel-heavy e-commerce, this reading is not about your cargo, and using it as a demand proxy imports a bias you will pay for later.

The survey covers one month of purchasing managers' answers. Orders convert to production in weeks, to inbound components in one to two months, and to outbound finished goods in two to four. So the freight consequence of this reading belongs to the next quarter, not to this one. Spot moves first, contract renewals follow three to six months later, and the 2027 negotiation will be priced off whatever the chart looks like in the fourth quarter, which is not necessarily what the dock looks like.

What the coverage has not separated is the sign of the delivery time component. Every summary I have seen lists lengthening supplier deliveries as evidence of strength and stops there. It is evidence of strength in the index and evidence of constraint in your inbound program, and the two readings imply opposite actions for the same buyer in the same month. There is a second omission: no index level is quoted, only the relative position, and relative position cannot substitute for a level when the dividing line is fixed. Coming back to the line from below and climbing further above it are different events, and the story is being told as though the second had happened.

Quantify what the delivery time line costs, with assumptions stated. Suppose a component you consume at 1,000 units a week has a quoted lead time that extends from 45 days to 50, a five-day extension I am assuming for the arithmetic and not a figure from the release. Covering five extra days at that run rate takes about 715 units of additional safety stock. On a part carried at 20 dollars with a 20 percent annual carrying cost, four dollars a unit a year, that is roughly 2,860 dollars a year for one part. Multiply by the twenty parts on your critical list and the lengthening delivery times line, which reads as good news in the index, is a six-figure inventory decision. The index never shows you that number.

If the lengthening is supply-side, meaning capacity was cut and the delivery line is measuring a bottleneck rather than demand, then the rise is partly a constraint reading, and treating it as a demand signal means locking capacity at the top of something that is not growing. If the order strength includes cargo pulled forward ahead of duty dates, the same reading precedes an air pocket. Neither possibility is visible in the headline, and both are checkable against your own data within a month. The boundary is this: the index gives direction and breadth in manufacturing. It does not give magnitude, it does not cover services or non-manufactured cargo, and it cannot separate demand from constraint. Past that boundary you are reading your own order book, and that is the only number that sizes anything.

↑ Back to the key points

— By Dr. Ingrid Voss