German and Dutch dock strikes hit Hamburg, Bremerhaven and Rotterdam

The Analysis

Six German ports, one bargaining round, forty-eight hours. Hamburg, Bremerhaven, Bremen, Wilhelmshaven, Emden, Brake. Add Rotterdam, Amsterdam and Zeeland on the fourth of September and you no longer have a port dispute. You have a corridor with its gates shut on the same afternoon.

It was a warning strike, called after the ver.di talks collapsed, and the carriers' note about it is more interesting than the dates: vessel bunching, and truck access suspended at APM Terminals MVII. Both phrases describe what happens after the gates reopen, which is the part that eventually costs you money.

Read only the dates and you will plan for two days. Read the carrier note and you will plan for two weeks. The distance between those two readings is where this piece lives.

A stopped berth is easy to picture. A berth window is harder, and the window is the thing that actually breaks. Windows are allocated against the proforma schedule, a document drawn up months ahead and sold to shippers as a promise. The window does not move because a ship is late; the ship moves to fit the window, or it loses the window to a ship that is on time. So when forty-eight hours of calls are displaced, the port does not work through a backlog in arrival order. It re-sorts. Re-sorting is where the cost hides, because the ship pushed to the back of the queue is chosen by volume and by the carrier's own network priorities, not by how long it has been waiting.

That is what bunching means, and it is worth being precise about why it costs more than the stoppage. Two days of lost gate throughput produces a queue whose unwind time is longer than two days, because port capacity is roughly fixed while the queue is not a line but a set of competing claims on the same cranes. Put the Dutch action on the same day and the overflow has nowhere to go: Rotterdam is not a relief valve when Rotterdam is also shut. The whole North European schedule then absorbs the shock at once, and schedules already described as tight have no reserve with which to absorb anything. A stoppage consumes time. Re-sequencing consumes the schedule, and the schedule is what your booking was priced against.

The physical signature of this sits at the gate, not the quay. Cranes run around the clock; gates run on shifts and appointments. When two sailings discharge inside one window, boxes reach the gate faster than the gate can process them, truck turn time doubles, and the queue that was on the water becomes a queue on the road. That is what the suspension of truck access at APM Terminals MVII is telling you, and it is not a detail about a single terminal. It is the point where a water-side problem turns into a landside one, and landside problems have a longer tail, because a truck that could not get in today still has to get in tomorrow and the day after.

Who pays depends on where you sit in the re-sort. Shippers holding fixed allocations on the big direct services are re-sorted last, which is another way of saying they are fine. Shippers on the smaller calls, or on services whose operator has a larger ship to protect, get moved. Importers holding promotional stock are hurt before importers holding safety stock, because a promotional window is a date and safety stock is a number. North European exporters feel it as a missed cut-off rather than a late arrival. Almost nobody in Asia feels it this week, which is why the story reads as a European story even though the schedule that breaks is a global one.

The part I have not seen argued anywhere is that a warning strike is deliberately cheaper per day than a real one and dearer in total. Two days is short enough that no carrier will omit a port over it. Omitting a port is a decision, decisions get communicated, and communicated decisions let shippers reroute. So nobody omits, everybody waits, and the waiting is what manufactures the bunching. Stretch the same action to eight days and the outcome for cargo would probably be better, because the uncertainty would resolve into a plan: omit Hamburg, discharge at Bremerhaven, truck it up. Four days of certainty cost less than two days of doubt. The instrument is built to impose cost without forcing a decision, and the cost it imposes is exactly the cost of not deciding.

Timing runs later than the headlines. The strike ends on the fourth. The gate queue peaks around the sixth to the eighth, when the first post-strike vessels discharge into an appointment book built for a normal week. Bunching shows up on the water over the following ten to fourteen days, as two calls arriving inside one window. The equipment effect, empties stranded inland because the return leg slipped, lands three to five weeks out. The schedule is not restored until the carrier either inserts a recovery call or eats a week somewhere else in the rotation, and that choice is usually taken at the next proforma review rather than this week.

Numbers, all mine and not the source's. Assume twenty-five FEU a week into North Europe, split sixty-forty between Hamburg and Rotterdam, with three weeks of cover at the destination. Assume the forty-eight hours costs three days of gate throughput once you count the ramp-down and the restart, that bunching adds five days on the water, and that double volume through the gate adds two days of turn time. That is ten days against twenty-one. Three of them are the strike. The other seven are re-sequencing, and re-sequencing is the part you can influence, because a gate slot booked before the vessel arrives is processed in the order it was booked rather than the order the boxes turned up.

If only Hamburg had stopped, Bremerhaven would have taken the volume and the re-sort would have stayed local, resolved inside a week. If the Dutch action had landed seven days later, Rotterdam would have been a relief valve instead of a second closed door. What made this expensive was not the size of either action. It was that both were drawn from the same two days, so the substitute and the original went down together.

The consequence is that you stop treating North Europe as one node. It is at least two labour markets, two road networks and two sets of terminals, and the value of an alternate is measured by how little it shares with the primary, not by how close it looks on a map. Closeness is precisely what makes an alternate useless in a week like this one.

The trap is measuring the stoppage and missing the re-sort. Forty-eight hours appears in every report; a schedule that has lost its reserve appears in none, and it will still be missing in November. The second trap is assuming a short action is a small one. Duration is the least informative variable here. What matters is whether the substitute set sat inside the same blast radius, and this week it did.

Six ports on one coast, one bargaining round, one afternoon. The map looks the way it did in August. The difference is that the redundancy you thought you owned was drawn from the same well as the risk, and a reserve that shares a source with the hazard is not a reserve, it is a second exposure. Ask each of your alternates what it shares with your primary. That answer decides whether the next notice costs you two days or two weeks.

↑ Back to the key points

— By Tomás Ferreira