Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- Put a number on the bet before the next purchase order: at an 8 per cent cost of capital, six extra weeks of holding costs about 0.9 per cent of goods value, so front-loading only clears when expected increase times probability of landing exceeds that.
- Split the SKU list this week into storable and seasonal, and front-load only the first group; the second needs a September capacity plan instead of an earlier sailing.
- Book September and October capacity now, ahead of the withdrawal that typically lands four to eight weeks after a volume cliff, and confirm it in writing rather than assuming the space rolls over.
- Cap the front-load at verified warehouse capacity rather than at appetite: inventory you cannot store is a markdown in the fourth quarter, not protection against a tariff.
- Set a decision date two weeks before the August date; if the expected increase has not been confirmed by then, stop pulling cargo forward and let the borrowed inventory drain.
The NRF/Hackett Global Port Tracker projected 2.47 million TEU of U.S. imports in July - the highest monthly total on record - as importers front-load ahead of an expected August tariff jump. Volumes are forecast to slide from August; LA/Long Beach reported almost no ships at anchor over the 4 July weekend despite the surge.
The Analysis
A record month and an empty anchorage, in the same report, in the same week. The NRF/Hackett Global Port Tracker projects 2.47 million TEU of U.S. imports in July, the highest monthly total on record, as importers pull cargo forward ahead of an expected August tariff increase. Volumes are forecast to slide from August. And over the Fourth of July weekend, Los Angeles and Long Beach reported almost no ships waiting at anchor.
That combination is the story, and it is not a story about things going well. Four years ago a record month produced a visible queue down the coast. This time the same kind of record produced nothing to look at. The constraint did not vanish; it moved upstream, into a schedule that now swallows the surge before it reaches American water.
So the useful question is not how large July was. It is what was borrowed to make it large, and who pays it back.
Put the three facts side by side and a property of front-loading becomes visible, one that rarely gets stated. Pulling cargo forward does not create volume. It relocates volume on the time axis. Whatever moved into July came out of August, September and October, and the bill for a borrowed month is always the slope that follows it.
Why the water was empty is worth spelling out, because the obvious reading is the wrong one. This wave was planned. Importers booked early, so arrivals spread across a month instead of landing in a fortnight, and terminal planning had something to plan against. The binding constraint also sat upstream, in berths and equipment and in a rotation that no longer closes, and a queue there is invisible from San Pedro Bay. Front-loading flattens the arrival curve at the destination by steepening it at the origin. The curve does not disappear. It changes address.
Front-loading is not available to everyone. High-value goods that store well and do not perish can be pulled forward. Fashion, fresh produce and anything with a season attached cannot: a fourth-quarter line cannot arrive in July and wait. That second group did not participate in the record and gains nothing from it, and it will still pay for it in September.
The timing sits here. An August slide does not turn into withdrawn capacity immediately. Carriers respond to a volume cliff with a lag of roughly four to eight weeks, which puts the capacity withdrawal in September and October, exactly when the cargo that could not be pulled forward needs to sail. The pain is not in July. It is in the quarter after July, and it arrives disguised as a booking problem.
Here is what the coverage left out, and it is the only number in this story that a shipper can actually use. The cost of front-loading is the carrying cost of holding inventory longer. Assume an 8 per cent annual cost of capital and six extra weeks of holding, and the cost comes to about 0.9 per cent of goods value. That is my assumption, not a figure from the report. Once you have that number, the decision stops being a feeling. Front-loading pays only if the expected increase, multiplied by the probability it actually lands, exceeds roughly one per cent of goods value.
If the increase slips, or comes in below that threshold, you have paid the carrying cost and bought nothing, while carrying six extra weeks of stock into a quarter where demand is softening. The second thing left out is smaller but useful: an empty anchorage is a leading indicator, not a clean bill of health. It points to an arrival trough and a service cut, several weeks out.
One more magnitude, to see the shape. Assume, again as my own assumption, that this round of front-loading borrowed 10 per cent of August to October volume. Ten per cent of 2.47 million TEU is about 247,000 TEU relocated into July. The size is not the alarming part. The shape is. A spike was bought with a hole, and the bottom of the hole falls in the same weeks the capacity withdrawal takes effect.
Two traps are worth naming before the next purchase order goes out. The first is treating front-loading as free insurance. It is not insurance; it is a position taken on a policy event, and it should be sized like a position, not like a plan. The second is counting only what goes into the warehouse and not what comes out of it. Borrowed inventory that does not sell is not a hedge, it is a writedown, and the writedown usually lands in the same quarter as the capacity problem. Reverse the outcome and the logic holds just as firmly: if the August increase never lands, the front-load is a pure loss, with carrying cost paid, no price advantage collected, and a pile of goods that has to be discounted in the fourth quarter. A world without front-loading would have produced an unremarkable July and a stable autumn. You traded a spike for a plateau, and a plateau is cheaper to plan against.
A port with no queue is not a port at ease. It is a port that has been relieved of a job it used to do, and the job went upstream, to a place nobody is watching. The map did not get simpler. It got longer.
↑ Back to the key points