Kuehne+Nagel: Manila worst in Asia at 4.69-day wait; Antwerp, Hamburg strained

The Analysis

Put the map on the table and three points light up at once, and they are not the three you would expect to be lit in the same week. Manila North is averaging 4.69 days of vessel waiting, the worst figure in Asia in Kuehne+Nagel's port update for 7 to 14 July, with vessels arriving outside their window delayed beyond five days and the MICT yard running at 85 to 90 percent utilization. Antwerp sits at 2.02 days with yard density described as critical, which is a heavier sentence than it looks. Hamburg is at 1.61 days, and the cause given is not berths or barges or labour but a lightning strike near Uelzen that interrupted rail service into the port.

Three numbers, three failure modes, and only one of them is about water. That is the sentence I would carry into next year, because by then the weekly figures will have changed and the structure underneath them will not have. What follows is how I read a report of this kind: not as a snapshot of a bad week, but as a reading taken from a network that has been quietly rearranging itself for a decade and has now told us where it is thin.

Manila's mechanism is the one that repeats. A vessel arrives outside its window and queues behind the calls that arrived inside theirs. A yard at 85 to 90 percent has between a tenth and a seventh of its slots free, and that is the entire shock absorber the terminal is working with. A box that cannot leave the yard fast enough becomes a vessel that cannot berth fast enough, and the queue grows from the land side outward rather than from the sea side inward. Which is why the wait time is the symptom and the yard is the mechanism, and why a plan built on the wait time is a plan built on last week's problem.

That distinction is not academic housekeeping. One number tells you what already happened to somebody else's vessel. The other tells you what is about to happen to yours, and it is published before your container is in any trouble at all. Port reporting rewards the first kind of number because it is dramatic, and shippers keep buying it because it is easier to quote to a customer than a utilization percentage.

The deeper problem in Manila is not quay length and rarely has been. A port of that kind is asked to do two jobs at once: serve a domestic distribution geography that is scattered across water, and act as a node in international strings. Both jobs discharge their boxes onto the same limited landside, and the landside is where the argument is settled. You can pour concrete at the water's edge for ten years and still come up short, because the constraint sits behind the gate, in the roads and the yards and the trucking that nobody photographs for the opening ceremony.

I have watched this pattern in enough places to be careful about predicting it away. Port capacity is a capital project with a ribbon and a date. Landside capacity is a hundred small decisions with no date at all, taken by people who do not think of themselves as running a port, which is exactly why the gap between the two tends to widen while everybody is announcing the good news.

Antwerp shows the same mechanism one stage earlier, and for that reason it is the more instructive of the two. Critical yard density is what a port looks like before the vessels start waiting, not after. The 2.02 days is a bill presented for a yard problem that began when boxes stopped leaving as fast as they arrived, and by the time it reaches a public average the cause is already old. A river port's yard is a schedule question more than a crane question. If the barges are late, if trucks are turned at the gate, if an inland terminal fills up, the boxes stay where they are and the density climbs, and none of that is visible from the sea. All of it eventually shows up as a vessel that waited.

Hamburg is a different animal entirely and I want to be precise about why, because the number looks like the Manila number and shares almost nothing with it. A lightning strike near Uelzen interrupted rail service into the port, and the 1.61 days that followed is a landside single-point failure, not berth congestion. The quay was fine. The water was fine. The boxes could not leave by the mode they were booked on, and the ship absorbed the difference by waiting at anchor with nowhere useful to go.

That is the sentence that ought to worry a European importer more than the Manila figure does. A port's published performance can now be moved by weather a long way inland, on a corridor that was laid out for a previous century's traffic and is being asked to carry this century's. When a port's number can be broken by a storm over a rail junction, the port is no longer a place. It is a junction in a longer line, and the line is what you are actually buying when you book.

Averages are where port reports go to die. The Manila average of 4.69 days blends a windowed vessel that waited a short while with an off-window vessel that waited beyond five, and if your booking sits on an off-window string then the average describes a port you are not using. The off-window figure is the one that belongs to your cargo, and it is the one the headline never carries. This is the quiet violence of a mean: it averages away precisely the population your shipment belongs to.

The European numbers mislead in a subtler way. Antwerp's 2.02 and Hamburg's 1.61 are averages for two ports whose cargo is not interchangeable, and reading them as a single verdict on European congestion produces the most expensive kind of comfort, the sense that the continent is mildly tight when in fact one gateway has a yard problem and the other has a rail problem, and the remedies have nothing in common. One asks you to slow your inland plan down. The other asks you to change your inland route. Confusing the two costs a week.

Layered by trade, the exposure separates cleanly. Philippine importers and exporters carry the worst of it, and inside their cargo the split runs between windowed and off-window services, which means the reliability of a booking is a property of the string and not of the port. Asia-Europe cargo through Antwerp feels density as gate time and truck turnaround rather than as vessel waiting, and the number it should be watching is not the one in the headline. Cargo into northern Europe through Hamburg feels the disruption inland, as days added to a rail leg that the port's own figure does not count and cannot count.

There is also a geography to Manila that the number flattens. It is a domestic distribution hub at the same time as it is a node on international strings, so a delay there is charged twice: once to the domestic leg that has no alternative, and once to the connection the box will miss at the next port. The second charge is invisible in a port report because it lands on somebody else's schedule, and the cargo that suffers most is the cargo moving on the tightest promise. That is usually the cargo nobody has flagged as urgent in the spreadsheet, because urgency is a property of the promise rather than of the commodity.

Run the buffer arithmetic on the Philippine lane and the exposure stops being abstract. Assume a shipper moving twenty containers a month, five to a sailing across four sailings. If each sailing lands five days outside plan, that is a hundred container-days of inventory sitting where it was not scheduled to sit, a little over three container-weeks of working capital parked in the wrong place at the wrong time. The volumes and the five-day variance are my assumptions, chosen so the arithmetic can be followed; the 4.69 average and the over-five-day off-window delay are the report's numbers, not mine.

Now the same lane at ten times the size. Two hundred containers a month, forty to a sailing across five sailings, five days late each time. That is a thousand container-days, thirty-odd container-weeks, and the figure stops being an inconvenience and becomes a line on a balance sheet that a treasurer will ask about. What changes at that scale is not the arithmetic. It is the menu. A shipper that size can buy its way out with a second gateway, a dedicated inland move, or tonnage of its own, and a shipper of the smaller size can do none of those things.

Which produces the quiet unfairness of congestion, and it is worth naming because it shapes which remedies are actually available to you. The small shipper's only real tool is time, so it buys time in the form of buffer and pays for it by promising its customer a later date. The large shipper buys optionality and pays for it as a line item. Both decisions are rational. Only one of them ever appears in a freight rate, which is why the market keeps arguing past itself about what congestion costs.

The cost of buffer deserves to be stated plainly, because people price it wrong almost every time. A buffer is not a freight cost. It is inventory, it is warehouse space, it is the working capital that sits in a box instead of a bank, and above all it is the difference between a date you can keep and a date you are hoping to keep. That difference does not show up in any carrier's quotation and it is the largest number in this article for most readers.

Most of the remedies a shipper actually has are annual rather than weekly, which is the part a report like this tends to hide. Contracts are bid once a year, minimum quantities are committed, inland haulage is priced against a schedule that assumes a port behaving normally. When a week like this one arrives, the only lever still in your hand is the one you bought last year, and if last year's contract assumed a fast gate and a reliable rail corridor, you will spend this week discovering what you did not buy. That is not a criticism of anybody's procurement. It is simply where the leverage sits in a business whose operating unit is a sailing and whose contracting unit is a year.

Step back from the week, because a report of this kind is a photograph and what we are trying to read is the drift. Over the last decade the industry added a great deal of capacity at sea and a much smaller amount ashore, and the imbalance was tolerable for as long as the network held slack enough to absorb it. The slack is what is being consumed now, in weekly increments, and a terminal at 85 to 90 percent yard utilization is a picture of a place where very little of it remains.

Ten years ago you could describe a shipper's routing as a lane with a port at each end, and the description held for most cargo. It no longer does. A box now touches a feeder port, a hub, a mainline port, an inland terminal and a warehouse, and the reliability of the whole journey is the product of the reliability of each part. That is a cruel arithmetic, because products of fractions get small quickly. Three nodes each running at ninety percent is not a ninety percent journey. It is seventy-three, and the missing twenty-seven points are distributed among people who never agreed to carry them.

There is a rhythm underneath all of this that explains why the same complaint returns every few years. A port takes most of a decade to plan, permit and build, and a trade lane can reorient itself inside a single season. Capacity therefore tends to arrive as the answer to a question that has already changed, which is why the gap between quay and hinterland does not close when the money is finally spent. The decade we are in is not unusual for having congestion. It is unusual for how little slack is left to absorb it, and slack is the one thing nobody puts in a business case, because it looks like waste right up until the week it is needed.

Which is why a single port's number is really a reading of a system rather than a report on a place. The same rotation that waits in Manila is scheduled to connect at a hub and discharge in Europe, and a delay at the first node propagates as a missed connection rather than as a longer voyage. Ports are where a network admits its problems. They are not where it creates all of them, and treating them as the source is how people end up fixing the wrong end of the line.

The report also measures ships, and shippers do not own ships. It counts waiting time at an anchorage, which is a cost the carrier absorbs first and passes on later, and it says nothing about the container that missed a production slot because a feeder was late or the promotion that ran out of stock because a rail corridor had a bad afternoon. The gap between what the report measures and what you actually lose is where your exposure lives, and no weekly update will ever close that gap on your behalf.

Manila's counterfactual is seasonal and unfriendly. If typhoon activity reaches the Philippines in the third quarter, waits worsen before they improve, and the over-five-day off-window delay stops being the tail and becomes the base case. Planning on 4.69 in that world is planning on a number that no longer exists, and the shipper holding a customer promise built on it will be the one explaining the gap.

Antwerp's counterfactual is a threshold rather than a season. Critical density has no cushion left in it, so any event that slows outflow further, labour action, equipment failure, an import surge arriving a few weeks early, does not degrade the port smoothly. It tips it, and the tipping spills into September, which is precisely the month when nobody in northern Europe has spare days to lend.

Hamburg's counterfactual runs the other direction and is the most misleading of the three. If the rail corridor is restored quickly, the port figure normalizes quickly, the chart looks recovered, and the backlog the interruption created is still working its way through inland transit for some time afterwards. The number recovers before your cargo does. In my experience that gap has cost more than the original event, because the original event is visible and the recovery is assumed.

What I would actually do, then, and I will keep it to things that survive the next weekly report. Build the buffer into the promise rather than into the expedite. On Philippine lanes the planning number is five to seven days, not 4.69, and a customer date that assumes the average is a date you will miss politely and repeatedly.

Read the yard rather than the wait. Yard utilization precedes waiting time, it is published weekly, and it gives a warning that no average can give because the average arrives after the damage. A terminal at 85 to 90 percent is a terminal where same-week pickup is an assumption rather than a plan, so book equipment and slots before you need them and treat a move above ninety as a signal worth a phone call rather than a curiosity for the file.

One reading of this report is a rumour. Eight consecutive readings are a trend, and the difference matters more than any single figure in the table. A port that shows 4.69 once had a bad fortnight. A port that shows a number in the fours for eight weeks has a structure, and structures are what you redesign your routing around, because they will still be there when the weather changes and the news has moved on to somewhere else entirely.

For Europe, keep the two halves of the problem apart. Antwerp cargo needs a weekly density reading and a truck and barge plan that does not assume a fast gate. Hamburg cargo needs an alternate inland routing quoted in advance, in writing, for the boxes where a missed date is expensive, because the day you need that routing is the day nobody on the other end picks up the telephone.

And then the part that takes a year rather than a week. Gateway diversification is a project, not a booking decision. It means qualifying a second port, a second inland corridor and a second set of documents, and paying for that capacity in the months when you do not need it, which is the only time it is ever available to you. Shippers who did that work before this July are having a dull month. Shippers who did not are having an expensive one, and the difference did not appear in any weekly report.

The map will look different in ten years and it will not look different because somebody redrew it. It will look different because a yard filled up, a rail corridor failed in a storm, a queue formed behind a window, and a few thousand decisions were taken in response to each. Ports are slow things, and the lines on the chart do not straighten on their own. The ones that matter most are the ones drawn inland, where nobody is taking a photograph.

↑ Back to the key points

— By Tomás Ferreira