Drewry World Container Index Rebounds 3% to $2,286 After Three-Week Slide

The Analysis

Three percent up, to 2,286 dollars per 40-foot container, in the week ending 26 June. That is the Drewry World Container Index, and the headline is not really the interesting part. The interesting part is the sentence right behind it: only 24 blank sailings are expected across the major east-west trades over the next five weeks, a cancellation rate that Drewry puts well below seasonal norms.

Read those two together and you get something slightly unusual. Rates went up while the fleet stayed in service. Most weeks where the index jerks upward, carriers have quietly parked a few ships to make it happen. Not this one. Twenty-four void sailings spread over five weeks is roughly 4.8 a week across every major east-west trade combined, which is not a capacity squeeze by any reading of the word. So the move came from the demand side, described here as resilient spot demand heading into July, with peak-season surcharges and general rate increases being filed on top.

Before anyone gets excited about a bottom, two things are missing from what I was handed. I do not have the size of the three weekly declines this print interrupted, so I cannot tell you whether one week of plus 3% recovers a tenth of the slide or half of it. And I have no lane-level breakdown at all this week, which means every lane-specific claim is off limits for me. What follows works only with the numbers above, plus arithmetic I show you, plus assumptions I state out loud.

Let me do the arithmetic that is available first, since that is where any real decision starts. If 2,286 is the level after a 3% rise, then dividing by 1.03 puts the previous week around 2,219, and the step itself was roughly 67 dollars per box. Sixty-seven dollars is not a freight crisis. It is also not nothing, and the trick with steps is that they compound in the places nobody watches.

Convert that step into the units your commercial team actually argues about. Assume 1,200 units in a box, which suits light assembly cargo: 67 dollars a box is about 0.06 dollars a unit. Assume instead you ship by volume and a 40ft gives you 67 usable cubic metres but your cargo really fills 58 of them, so the effective cube is 58: that same step is 1.15 dollars per effective cubic metre per week. Whether that matters depends entirely on what sits on either side of it, which is the next question.

Here is what I cannot tell you, and I would rather say it than pad the column out. I do not know the depth of the three weekly declines that preceded this print, because the figure was not given. If those weeks were shallow, then one plus week means more than it looks. If they were steep, then this 3% barely registers. Ask whoever sends you the weekly note for the four-week series before anyone says the word bottom in a meeting. A direction change is one observation. It is not evidence of anything yet, and treating it as evidence is how budgets get rebuilt twice in six weeks.

I also cannot convert those 24 blank sailings into a percentage of capacity, because nobody handed me the denominator. Cancellations get quoted as a rate against the scheduled programme, and I do not have the scheduled programme. So when Drewry calls the rate well below seasonal norms, I am repeating a qualitative judgement, not converting it into a number you can model. Do not let anyone in your building turn it into one.

What I can do is price the two directions. Assume you ship ten boxes a week across the east-west trades, which is about forty-three a month and fifty over this five-week window. Take the step of roughly 67 dollars and let it repeat each week in the same direction. If it repeats upward, the level walks 2,286, 2,353, 2,419, 2,486, 2,552 and 2,619 by the end of the window. If it repeats downward, it walks back through 2,219, 2,153, 2,086, 2,020 to 1,953. The distance between those two end points is 666 dollars a box. Across fifty boxes, that is 33,300 dollars of difference riding on a move everybody in your company will describe as three percent.

Compounded rather than added, which is the more honest way to read a percentage series, your endpoints are about 2,650 up and about 1,972 down, a spread of 678 dollars. It barely changes the point. The point is that a number small enough to be mentioned in passing on a Monday morning is worth a five-figure swing over one planning window on an ordinary mid-sized book.

One practical use of the level itself, and then I will stop quoting 2,286. Use it as a reconciliation exercise rather than a benchmark. Take one real all-in quote from last week on one east-west lane and split it into three buckets: the ocean component, every filed surcharge, and everything that is not ocean freight at all, meaning origin and destination charges, haulage and whatever else travelled into the same quotation. If the ocean component sits far above 2,286, something specific is causing it and you want to know which bucket is doing the work. My experience is that the divergence usually sits in the third bucket, which is also the bucket nobody negotiates. Do this for one quote this week rather than for all of them. One clean reconciliation teaches you the structure; fifty done badly teach you nothing, and the whole point is that the index is not your price. It is a reference for asking why your price looks the way it does.

Now the capacity read, which is where the real content of this week sits. Twenty-four blank sailings over five weeks works out to 4.8 per week across all the major east-west trades put together. That aggregate is not your programme. Some of your services will carry no void sailings at all; one of them may lose two weeks in a row. Arguing about your exposure from a market-wide average is how people end up shut out of a sailing they assumed was running. The count tells you the market tone. It does not tell you anything about the specific vessel your container is supposed to be on, and only your carrier can tell you that.

For space, the direction of this reading is kind. A low cancellation count means sailings mostly run as advertised, which lowers the risk of being shut out or rolled for want of space. You do not need to chase space you do not have cargo for, and you do not need to book three weeks earlier than usual out of fear. The anxiety premium that a lot of shippers pay in peak season is a payment against rolled bookings, and right now there is not much evidence that rolled bookings are the problem you should be insuring against.

For price, the reading cuts the other way, and this is the part worth writing down. When carriers pull capacity, they are building a floor under rates: if demand softens, they can simply pull more. They have done none of that here. So this rebound has support from demand but no structural floor beneath it. If the spot demand described here proves less resilient than hoped in the second half of July, there is no parked tonnage standing between your freight rate and the ground. That asymmetry is the actual story this week.

Put the calendar next to it, because five weeks is a finite and named period. Counting from the week ending 26 June, the window covers the weeks ending 3 July, 10 July, 17 July, 24 July and 31 July. That is your booking horizon for this assessment. Every box you intend to move inside it is already known to your planning team, or should be, because these are July shipments and it is now the end of June. Nothing about this index changes what you ship. It changes what you pay, and possibly when you commit.

Which raises the question I get asked every time an index turns: should we wait and see whether the slide resumes. Let us price that instead of guessing. Assume the goods in one box carry an invoice value of 50,000 dollars and your cost of carrying inventory runs 8 percent a year. Weekly carrying cost is 50,000 times 0.08 divided by 52, which is 76.92 dollars a week per box. Your prize for waiting one week is at most one step, 67 dollars, and only if the market falls in that direction. So on those assumptions waiting loses about 10 dollars a box per week at best, before you consider any service consequences.

Lower the assumed value to 30,000 dollars and weekly carrying cost falls to 46.15, so waiting becomes cheap. Raise it to 80,000 and weekly carrying cost is 123.08, so waiting gets expensive fast. The sharper way to state it: at the 50,000 dollar assumption, waiting pays only if you expect the index to fall by more than 76.92 dollars a box in a single week, which is 3.4 percent, which is faster than it rose this week. You would have to believe the slide accelerates to justify sitting still. I do not have the size of the prior declines, so I cannot tell you whether that belief is reasonable, but whoever is arguing for waiting should be made to state that expectation as a number.

Then there is the committing question, and here the small size of the step works in your favour. Take the same ten boxes a week, fifty over the window, and assume half the book is fixed at today's 2,286 while half rides spot. In the up case, spot over the five weeks runs 2,353, 2,419, 2,486, 2,552 and 2,619, averaging 2,485.80, so the half riding spot pays 199.80 a box more than the fixed half. In the down case the same five weeks average 2,086.20, so the fixed half pays 199.80 a box above what spot would have cost. Across fifty boxes each outcome lands within about 9,990 dollars of the other, against roughly 114,300 dollars of freight if every box had cleared at today's level.

That symmetry is the useful finding. When the weekly step is this small, the cost of being wrong is smaller than the cost of the meetings most companies convene to avoid being wrong. Split the book, write the ratio down and stop. If you would rather not revisit it until September, this is one of the rare weeks where that laziness is close to free.

And here are the four things I still lack, which I am listing because leaving a blank in a plan beats filling it with a guess. The depth of the three weekly declines before this print: ask whoever circulates the weekly note for the four-week series, owner is whoever maintains your rate file, and nothing further in this column belongs in a plan until it arrives. The total scheduled sailings standing behind the 24 cancellations: without the denominator there is no cancellation percentage to argue about, so stop letting that one word do work it cannot do. Any lane-level detail for this week: I have none, which is why nothing above claims to describe what your particular service did. And the actual size of the peak-season surcharges and general rate increases on your lanes: no formula rescues you from not knowing those figures, and every scenario here changes shape the moment they land. Four blanks, four people who can close them, none of them you guessing at your desk.

Then there is the layer problem, which never gets the attention it deserves. What Drewry published is a composite freight measure per 40-foot container. What carriers are doing, per the same note, is layering peak-season surcharges and general rate increases on top of the base. Those are two different documents. Your invoice reads base plus whatever has been filed, and nothing guarantees the two move together.

Run it as a formula rather than a speech. Call your base movement B and any additional filings landing in the same week S. Your invoice moved by B plus S, while the headline said B. If B is around 67 dollars this week and some combination of filings adds another 200 dollars a box, the index explains 67 out of 267, which is 25 percent of your increase. If those filings add 100 dollars instead, the index explains 40 percent. If they add nothing on your particular contract, the index explains everything. There is no version where you can skip asking which applies to you.

That leads directly to the paperwork, which is where the money usually hides. Get the surcharge stack in writing: each charge, its amount per box, its effective date, how much notice you get before it changes, and whether it sits inside or outside the all-in figure you were quoted. If a contract adopts carrier surcharges by reference, then a flat index week tells you nothing about your exposure and the base rate becomes decorative. Two practical asks here. Any charge that can be revised without written notice should be capped before you renew. And every quote from now through July should arrive split into two lines, base and everything else, because a single blended number makes it impossible to see which part of your cost is moving and why.

On demand itself, the note says spot demand heading into July is resilient and that is doing real work in this analysis. Resilient demand, alongside rising filings, means the risk you are managing is not a collapse in volumes that leaves you holding stock. It is the ordinary risk of paying more next month than this month, plus the less ordinary risk of the filings stack growing faster than the base. Those two risks have different remedies. The first is answered by committing earlier or hedging part of the book. The second is only answered by contract language, because no amount of clever booking beats a surcharge clause that lets charges move without asking you.

Watch one thing rather than acting on it. Twenty-four blank sailings is this week's expectation, and expectations get revised. If the count climbs over the next few assessments, the market changes character: space becomes the scarce thing, your rolled-bookings risk comes back, and every argument above about not paying an anxiety premium goes in the bin. Checking one number takes minutes. Being surprised by it costs a quarter.

The mechanics worth re-checking while you are in the invoices anyway are boring and reliable. All-in definitions differ between carriers and between offices of the same carrier, so get yours written down with the components named. Free time and demurrage tend to look generous until volumes rise and boxes start sitting, and a five-week window with resilient demand is exactly the shape of period where terminal dwell quietly extends. Ask what the commitment actually is when a sailing does not run: replacement vessel, next available sailing at the booked rate, or your booking simply repriced at whatever the week brings. None of this is exotic, and none of it is the sort of thing anybody volunteers.

On the budget question, here is my rule and it has saved me more embarrassment than any clever analysis. One weekly print does not move the plan. Two consecutive prints in the same direction on your own lanes does. This week gives you a plus 3% on a composite with no lane detail, and that is not enough to rebuild anything. Mark it, note the direction, and wait for the second observation. If the following assessment rises again, revisit the assumption. If it falls back, you have spent nothing except attention, which is the cheapest input you have.

I would apply the same discipline to how you read the drivers. Resilient spot demand and a light cancellation programme are both real, both cited, and both could be gone in two weeks. Neither of them justifies extending an inventory build beyond documented demand, and neither justifies cutting one either. What they justify is a re-check scheduled for a specific date, owned by a specific person, rather than a general feeling that things are turning.

Concretely, then. Today is 29 June, the day this landed. Tuesday 30 June is for getting the void-sailing calendar out of every carrier you use on the east-west trades, in writing, across the five weeks ending 3, 10, 17, 24 and 31 July. Wednesday 1 July is for the surcharge schedule with amounts, effective dates and notice periods. Friday 3 July is for reissued quotes split into base and everything else. Each of those has an owner, and each is a request you can send without knowing what next week's index will do, which is the point.

Every figure above is either in the note or derived from it in front of you: 2,286, roughly 67 dollars of step, twenty-four void sailings over five weeks, about 4.8 a week, and spreads of around 666 to 678 dollars between the two five-week endpoints. What is not there, I did not invent. Three weeks of decline followed by one week of increase is a coin toss wearing a narrative. Give it one more print before you buy it dinner.

↑ Back to the key points

— By Vivian Zhao