Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- Within one week, replace any buffer sized on the 14-day figure with one sized on average dwell, and cap added inventory days at 3 until your carrier reports a loop transit longer than its 12-month mean.
- Pull the last 12 months of your own booking data, compute average port dwell per gateway, and reprice the buffer at that number; a three-day average against a fourteen-day planning assumption releases 11 days of working capital per box.
- Before diverting any volume to Khor Fakkan, Fujairah or Sohar, obtain the current berthing queue in days from the receiving agent in writing, and divert only where that queue is at least 4 days shorter than the Panama wait.
- Add a congestion clause to every service contract at the next renewal: if average dwell at the contracted gateway exceeds 5 days for two consecutive months, the detention free time extends from 5 to 8 days automatically.
- Model the demurrage line at $150 per box per day and the carrying line at $13 per box per day, and present them separately to finance; the 11 to 1 ratio is the argument for paying for reliability rather than inventory.
Around 5% of global deep-sea container capacity - roughly 1.7 million TEU - is stalled in port delays, with acute bottlenecks across Asia, Northern Europe and the Middle East. Panama waits stretch to 10 days and alternative Gulf gateways such as Khor Fakkan, Fujairah and Sohar face berthing delays of up to 14 days, pushing shippers toward longer buffers and higher demurrage risk.
The Analysis
Five percent is a small number until you establish what it is five percent of. The release gives you the numerator and leaves the denominator to arithmetic: 1.7 million TEU at 5% implies a base of 34 million TEU of deep-sea capacity. That is the fleet the percentage is measured against, and it is the largest figure in the story.
Around that share of global deep-sea container capacity is stalled in port delays. The bottlenecks are concentrated in Asia, Northern Europe and the Middle East. Panama waits run to 10 days. The Gulf gateways that shippers have been told to divert to, Khor Fakkan, Fujairah and Sohar, are showing berthing delays of up to 14 days. Shippers are being pushed toward longer buffers and higher demurrage risk.
Two of those sentences describe a stock. Two describe a tail. Multiplying them together is the mistake I want to take apart.
A stock is a quantity at a moment: 1.7 million TEU is in the water right now, waiting. A tail is the far end of a distribution: 14 days is the worst berthing queue at three specific Gulf ports. They sit in the same paragraph, in the same tone, and every reader multiplies them, because 1.7 million times fourteen feels like a rate forecast.
It is not. Divide instead. If all 1.7 million TEU waited the full 14 days, the system would be turning over 121,000 TEU a day through those delays. That is 1.7 million divided by 14, and it is implausibly high for a delay-driven queue. Which tells you something usable: the average wait embedded in the 5% must be far shorter than the 14-day tail. If global throughput is around half a million TEU a day, which is my assumption from an annual volume near 180 million TEU, then 1.7 million TEU in the water corresponds to an average wait of about three and a half days.
Three and a half days is a different problem from fourteen. One is a schedule reliability issue. The other is a cash issue, and they are solved with different budgets.
Where does the 5% go? It does not vanish. Capacity in a queue is capacity that cannot make the next voyage, and its effect on the market is identical to capacity that has been laid up, with one difference: nobody announced it.
That difference has a contract consequence. A blank sailing is a countable event. It has a voyage number, a notice period, and usually a clause. Congestion absorption has no voyage number. A meaningful slice of the fleet can be eaten by queues while the capacity report shows the fleet unchanged. If your contract carries a capacity-withdrawal trigger, one that lets you reopen the rate or claim when the carrier pulls a stated share of capacity, congestion will never trip it, even though the economic effect is the same as the blank sailing that would.
Panama at 10 days and the Gulf at 14 days are not the same species of delay, and the difference decides what you can still do about it. Panama is a transit constraint. The queue is physical and ordered, there is a booking system, and the wait is largely a function of which slot you hold. Ten days there is a scheduling fact you can plan against, because the slot count gets published.
The Gulf ports are a berth constraint whose alternative is not an alternative. Khor Fakkan, Fujairah and Sohar are being recommended as relief valves for the same traffic, and all three are showing delays up to 14 days. When the diversion target is as congested as the origin, diversion stops being a hedge and becomes a second queue. The thing to check before you re-route is not the sailing distance to the alternative. It is the berth queue at the alternative.
The share of capacity tied up is a market number. What lands on a shipper is a per-box number, and it arrives in two lines that are usually kept in different folders. Demurrage is the visible one. Inventory carrying cost is the invisible one. Take a 40ft box of goods declared at $60,000 and a carrying cost of 8% a year. Both of those are my assumptions, not figures from the release. Eight percent of $60,000 is $4,800 a year, about $13 a day. Three and a half extra days is $46 of carrying cost. Fourteen days is $184.
Put a demurrage rate beside it. Even at a modest $150 a box a day after free time, fourteen days is $2,100, more than ten times the carrying cost of the same delay. The buffer everyone is being told to build is therefore not protecting service. It is protecting cash, and it is doing that at a ratio near eleven to one. Any buffer business case that nets the two lines together is understating the demurrage leg by an order of magnitude.
When does this reach you? Earlier than the delay does. Port congestion enters the P&L in a fixed order: demurrage invoice, then detention, then the rebooking premium, then the missed promotion. The first two are billed within a month of the event. The last two land a quarter later, when the sales number misses and nobody connects it to a berth queue in the Gulf.
The other timing fact is that buffers are self-defeating at scale. Every shipper who adds three days of buffer adds three days of inventory to the same system, and the ports do not get faster because the cargo arrived early. A buffer buys reliability with working capital, and it buys it only for as long as most of your competitors are not doing the same thing.
Here is what the coverage does not do with these numbers. It treats 1.7 million TEU as a capacity loss, which is correct, and then treats 10 and 14 days as the duration of that loss, which is not. The two figures describe different parts of one distribution, and the release never says which is which.
The missing number is average dwell, and it is the only one you need in order to size a buffer. It can be approximated from the two that were given, working from your own volume rather than the global one. If your carrier's Asia loop is turning in 70 days instead of 63, those seven days are in the water, and that estimate of your exposure is worth more than a 14-day headline from three Gulf ports you may never use.
The other missing item is the stability of the denominator. Five percent of 34 million TEU is 1.7 million. If the fleet grows while absolute congestion holds at 1.7 million, the share falls and the rate effect falls with it, without a single berth getting faster. Share is the number that prices. TEU is the number that gets quoted.
Had the release led with average dwell instead of the tail, the buffer decision would look different. A three-and-a-half-day average justifies three days and no more. A fourteen-day tail justifies a fortnight, which is 4.3 times the working capital for the same nominal service level. Same facts, four times the inventory.
The inverse is equally true. Size to the average, and if your cargo lands in the tail because it routes through a Gulf gateway or transships at a congested hub, you will be short exactly when the demurrage clock starts.
The trap is reading a stock and a tail as though they were a rate and a duration. One is a photograph of the system on a Tuesday. The other is the worst day of the year at three ports.
Size the buffer to the average. Price the tail separately. And check the berth queue at the port you are diverting to before you divert.
↑ Back to the key points