US Freight Volume Index Falls 9.3% Year on Year as Tariff Pull-Forward Unwinds

The Analysis

Let me start where the buyer actually stands, because nobody in these index stories ever does. My autumn book is already bought. My trailer and container commitments for August through October are largely fixed, because back in February I decided to get in front of the tariff deadlines and moved a large slice of my seasonal order forward. It was the right call at the time and I would make it again. But it carries a cost on my side of the table that a freight index never shows: I am carrying inventory earlier than my sell-through wants it, and inventory is cash I cannot spend on anything else. My payment terms with suppliers did not change, so I paid early on the goods and I am paying the freight on top of that, while my own customers keep paying me on the same 60 and 90 day clock they always did. What the market calls a freight event is, on my books, a working capital event.

Now the two numbers, read together. The US freight volume index for early August sits at 1.017, down 9.3% from a year earlier and 1.5% from July. Freight spend, the money the market paid out to move that cargo, sits at an index reading of 3.135, down 0.4% on the year and 2.8% on the month. Volume fell nine point three percent. Spend fell four tenths of one percent. If you shipped nine and a half percent less and paid almost the same, your cost per unit of freight went up, not down. The headline saying spend eased 0.4% is technically true and practically misleading. I do not read this as a carrier earnings story. It is a line in my landed cost, and this is money that ultimately comes out of my pocket.

So I want to do what I always do with a freight number. Put it on my side of the table, take it apart line by line, and ask what it changes about the next contract I am going to sign. The analysts quoted here say truckload, LTL and intermodal rates will only stabilise later in 2026 as the tariff pull-forward fades. That gives me a window, and windows close.

I should be honest about my own scorecard, because that is what actually decides whether any of this changes how I behave. I am measured this year on three things: landed cost per unit against budget, inventory turns, and in-stock rate on the categories my customers reorder. Freight volume is nowhere on my scorecard. So a headline saying volumes fell 9.3% does not cheer me up, it makes me look at the two lines behind it, because falling volume plus flat spend is the one combination that lands squarely on the number I am actually judged on.

Freight spend and freight volume are two separate indices, and the print does not tell us they share a common base, so I will not pretend I can just divide one by the other and call it a rate. But I do not need a perfect ratio to read a direction. I need to know whether my cost line is moving with the volume line or against it. Against last year, it is moving against it. Against last month, it finally moved with it, and that is the first honest crack of relief in this whole story. Let me put a number on that so you can argue with me instead of nodding along.

Assume, for the calculation only, that the volume index and the spend index share the same base period, which is the friendliest assumption I can hand them and probably more than is warranted. Under that assumption, the implied cost of moving a unit of freight is the spend reading divided by the volume reading. Year on year that is 0.996 divided by 0.907, about 1.098, so even with spend down 0.4%, the cost of moving a given unit of freight is up roughly 9.8% against a year ago. Run the same division on the monthly figures, 0.972 over 0.985, and you get about 0.987, so month on month that unit cost eased roughly 1.3%. Two completely different pictures out of one press release. Expensive against last year. Only just beginning to soften against last month. If your budget was built on the annual figure and your negotiation is timed on the monthly one, you are working to two different clocks and one of them will surprise you.

This is where the buyer's leverage either shows up or it does not, so let me be blunt. Carriers are looking at demand down 9.3% against costs that the same report calls sticky. Sticky costs mean a carrier cannot simply cut its way to a lower price; its break-even is propped up by fuel, labour, equipment and debt service that do not fall when volumes do. That is precisely the moment a carrier defends its rate floor hardest and competes hardest, quietly, for whatever volume is still moving. For me that is not a reason to celebrate, it is a reason to move. The shipper who opens a lane while the incumbent is nervous about empty equipment gets a very different conversation from the one who waits for the autumn peak and hopes for the best. The window this report describes is not that freight is cheap now. It is that for the next few months, the balance of nerve is on my side.

When I say I take a freight number apart, this is what I mean. My landed cost on an imported unit has several moving parts underneath the visible price: the base freight rate, the accessorials that only appear on the invoice, the inland leg at the destination, and the inventory carrying cost of how long the goods sit before they sell. A soft volume market pressures the first of those and, if I am careless, quietly inflates the second. Everyone watches the base rate, negotiates it hard and signs off on it, and then the accessorial line comes back at month end and eats the concession they thought they had won.

Before I renegotiate anything, I want a full twelve months of accessorial and dwell history on every lane I own, sorted by charge code, so I can see exactly what the headline rate leaves out. If I cannot see the charge codes, I cannot price the contract, and this is the part where I genuinely do not trust the first number I am handed.

Now the piece the market chatter keeps getting half right. The headline blames the volume drop on the tariff pull-forward unwinding, and that is fair as far as it goes: orders were pulled into the spring, the cargo moved, and the volumes are thin on the far side of that pull. What the headline does not say is that the pulled-forward cargo did not vanish when it moved, it handed itself to the buyer as inventory. My warehouse is heavier than it should be for this point in the season. My safety stock days are above target.

Every extra day of cover is cash, storage, insurance and the slow risk of a season that turns while the goods sit there. The freight market is soft partly because buyers like me are already full. So the same event that hands me negotiating leverage on the freight rate is quietly taxing me on the inventory line, and that trade only works if I use the freight window to lighten the stock as well, not just to win back a few dollars a box.

How long does a soft spot market take to reach a contract? Longer than a spot chart implies, and this is where a lot of buyers mis-time their ask. The report points to truckload, LTL and intermodal rates stabilising only later in 2026, which is analyst shorthand for soft through the near term and firmer once the pull-forward has fully washed out of the year on year comparison. In practice that means my leverage peaks before the market turns, not after. If I open a tender in the next few weeks I am negotiating into a soft market and I can lock something worth locking. If I wait for confirmation that rates have bottomed, I will be signing into the turn alongside everyone else who waited for the same signal. Leverage is a perishable good. It does not keep, and it does not come back at the same price.

Second piece of arithmetic, with the assumption on the table again. Suppose I hold safety stock on this category at 45 days of cover, and the front-load has pushed my actual cover to 60 days, so 15 days of stock I never planned to carry. Assume the annual landed cost of goods for the category runs 120 million dollars, and carrying cost, meaning storage, insurance, the cost of the tied-up cash and the risk of markdown, runs at about 12% a year on the value of the stock. Fifteen extra days on 120 million is 15 divided by 365 times 120 million, about 4.93 million dollars of goods held longer than planned, and 12% of that is roughly 592,000 dollars of carrying cost to pay for a timing decision I made back in February.

Now set that against what a freight concession is really worth. Take a lane where I move 200 containers a year and I win 150 dollars a box in the renegotiation, and that is 30,000 dollars. The inventory cost of the front-load dwarfs the freight saving unless I also sell the stock down. Any buyer who renegotiates the rate and forgets the inventory has optimised the small number and ignored the big one.

Who does this bite, and how hard? If you are a large importer with a freight desk and long-leased equipment, the soft market is a real gift and you have the people to take it. If you are a mid-sized seller buying truckload and LTL on the spot market, the same softness looks like a friendlier quote this week and a nasty surprise next quarter, because you are the one who funded the front-loaded inventory with more expensive money and you have the least room to wait. And the smaller you are, the less of this you can see at all, because you buy your freight as a line inside a package and never see the charge codes underneath it. Soft markets are not evenly distributed. They reward whoever is set up to read them.

Which categories matter here, because a 9.3% index is an average and averages hide everything. The categories that front-loaded hardest for tariff reasons are the ones now sitting on the heaviest stock, typically seasonal goods and anything with a long enough shelf life that buying early was a rational bet. If your category is one of those, your inventory overhang is the real story of this print, not the freight rate. If your category resists front-loading, because it is perishable or fashion-sensitive or simply short cycle, then you probably did not over-buy and this soft market is close to pure upside for you. Same index, two completely different outcomes, decided by something the index cannot see.

So how does the buyer's position actually shift, and what do I do with it? Three dials, and I turn all three. Contract tenure: I want shorter, not longer, because I believe the market is soft now and I do not want to be locked into today's number if it softens further, which is the exact opposite of the instinct to lock in while it is cheap. Index linkage: where I can, I want the rate tied to a published index with a collar, so I am exposed to the market rather than to the carrier's narrative about the market.

Volume commitment: this is the concession that costs nothing on paper and everything in practice. I would rather promise a percentage of my volume to a carrier than a fixed minimum quantity, and I want any shortfall penalty capped and spelled out, because in a soft market the most valuable thing I own is the freedom to move. The carrier wants my volume. I want their flexibility. Those two wants are tradeable, and this is the quarter to trade them.

Now the alternatives and the traps, because every one of them has a price on the other side. The lane itself is a variable in the same arithmetic, and I do not pretend to read a port the way Ömer Kaya does. He treats a gateway change as a cost change first and a geography question second, and he would tell me the saving on the rate line can disappear in the inland leg. So I keep the lane call with him and the money call with me, and I never let the two get mixed into one negotiation. A cheaper contract with longer tenure buys a lower headline rate and sells my future options; if the market softens further into the autumn, I am sitting on a rate I cannot reopen.

Going all spot saves the commitment and leaves me naked at the exact moment the pull-forward washes out and rates firm, and the report literally tells us that turn is coming later in 2026, so a pure spot book is a bet against a timeline we have already been handed. Splitting a modest volume across three carriers looks like leverage and is really a service risk, because being one carrier's real account buys priority that three small accounts never do. And the cheapest trap of all sits on the invoice: the accessorials, the detention and demurrage clock, the definition of free time, and whose yard the equipment sits in over a weekend.

A rate negotiation that leaves those lines untouched can hand back everything the headline rate gained. I have watched a buyer celebrate a five percent base rate cut and lose it in fees inside two months. On the equipment and dwell detail, the operational reading of that market is not my trade, and Ömer reads it far better than I do; my job is the contract, not the chassis.

Here is what I would put in motion, with dates, owners and numbers attached, because a plan without those is a wish. By Friday 14 August 2026, the category team must pull twelve months of accessorial and detention history on every lane above 100,000 dollars of annual spend into a single sheet, so the true cost of each lane is visible before a single conversation starts. By 22 August 2026, procurement must run a spot-versus-contract comparison on the three biggest lanes and set a written threshold: if the spot rate runs more than 8% below the incumbent contract rate for two consecutive weeks, we move at least a quarter of that lane's volume to spot and tell the incumbent we have done it.

By 29 August 2026, every lane over the threshold gets a tender letter asking for three specific things, a shorter tenure, an index-linked price with a collar, and a volume commitment expressed as a percentage rather than a fixed minimum, because the carrier's answer to those three tells us more than the rate number does. Through September 2026, the inventory owner must bring safety stock cover back to the 45 day target within eight weeks by holding new orders on the slowest moving 20% of SKUs, and must report the carrying cost saved against the arithmetic above.

And by 30 September 2026, before any rate is signed, the landed cost model has to show the freight change flowing through at no more than 4% above the current budget on any single lane, signed off by me. If a lane cannot meet that, it does not get signed. It gets re-tendered.

Let me argue the other side, because I do not want to be the buyer who is confidently wrong. If the pull-forward has not finished washing out, and importers still waiting on late tariff decisions push one more wave through late summer, then volumes stop falling, carriers get their pricing power back sooner than the report expects, and a buyer who went short and spot-heavy is the one caught without cover. That is the real risk in my own plan. The collar on the index-linked contracts is what protects me against it: it caps how far the rate can run away from us, and in exchange I give up the very bottom of the downside. I would rather own a capped way to be wrong than an uncapped way to be right. That is not caution for its own sake, it is arithmetic.

So if I were signing today, here is how I would do it. I would sign nothing long. I would take six month tenure, an index link and a collar, on a percentage volume commitment I can actually defend, on every lane where I hold real volume and the incumbent is more afraid of losing me than I am of losing them. I would spend this window on the invoice, not the headline, because that is where the money actually hides. And I would use the softness to sell down the inventory I built in February, because the freight saving is a rounding error next to what that stock is costing me every day it sits. The carriers are right that their costs are sticky. That is exactly why the buyer who moves now, and moves on the structure rather than the price, will be the one still standing when rates stabilise later in 2026.

↑ Back to the key points

— By Grace Yang