Panama Canal Restores Last-Minute Bookings at 33 Daily Slots After Record 521m Ton Year

The Analysis

Advisory A-39-2026 went out on 21 October 2026 and the headline word was restored: last-minute Panamax reservations are back.

Read one line further and the same document warns that the water deficit persists, that only the regular Transit Reservation system guarantees a transit date, and that unbooked vessels may face delays described as extremely extensive.

What came back is not capacity. What came back is the right to queue. We do the arithmetic on the two constraints that actually bind: one foot of draft, and thirty-three slots a day against 38.2 transits a day.

Start with the numbers that the canal itself published, because they are unusually rich. Fiscal 2026 closed with 13,951 transits and a record 521 million PC/UMS tons, past fiscal 2022's 518.8 million. Revenue rose 16.6% year on year. The hydrology behind it: rainfall 34% below average, with canal inflows down 44%. The mechanical response was staged: Panamax booking slots cut to 23 from 15 September, Neopanamax restored to 10 from 15 October, thirty-three slots a day in total, at 49 ft draft against a 50 ft norm.

Three of those figures deserve to be divided by each other before anything else is said. One: tonnage grew 521.0 against 518.8, which is 0.42%. Two: revenue grew 16.6%. Divide one compounded growth by the other and revenue per ton rose 16.1%. The record tonnage was nearly flat while the canal took materially more money per unit of cargo. If you moved boxes or grain through Panama in fiscal 2026, almost none of that extra revenue was you shipping more; it was you paying more per ton, through tolls, booking fees and slot premiums. A canal does not publish a 16.6% revenue line without somebody's cost per ton going up by roughly the same.

Two daily figures. 13,951 transits across 365 days is 38.2 movements a day. The current bookable slot count is 33. So the slot window covers 86.3% of the traffic the canal actually handled through the year. This is not a queue with slack in it. Even at record throughput, more vessels moved than there were slots to reserve, which means everyone else travelled outside the regular window and took whatever date they were given. Add the restored last-minute bookings and you have more tickets printed, not more seats.

Then the composition, which nobody reading the headline will notice. Thirty-three is 23 plus 10, and those two numbers serve two different fleets. Panamax is dry bulk, grain, some chemicals and smaller container tonnage. Neopanamax is where the big box ships run Asia to the US East Coast, and the entire Neopanamax last-minute pool gets ten slots a day. Ten. For every container line that failed to pre-book a Neopanamax slot on any given day, there are ten tickets, and the rest of that fleet is competing for them or waiting. Most coverage of this advisory will quote thirty-three and move on, and thirty-three is the wrong number for a liner operator.

Average payload per transit sits at 521,000,000 over 13,951, which is 37,344 tons per movement. Stated plainly, that is the scale at which a single transit decision is taken. A day lost at this unit size is not a schedule nuisance, it is a line item.

The hydrology is the part worth memorising, because it is asymmetric and it governs the next two quarters. Rainfall off 34%, inflows off 44%. Divide one by the other and you get 1.29: each further percentage point of rainfall deficit is producing about 1.3 points of inflow deficit. Groundwater does not run linearly into the lakes, and the gap between those two percentages is the measure of that. Which also means a forecast that says rainfall is heading back to normal should not be read slot for slot into restored draft. Assume roughly 1.3x amplification in either direction until the lake numbers say otherwise.

And that is why the word restored needs the quotation marks the canal itself put around it. A-39-2026 restores the ability to make a last-minute Panamax reservation while simultaneously stating that the water deficit persists and that only a regular Transit Reservation secures a transit date. Those two sentences are not contradictory, but together they say something quite specific: you may book now, and you still may not sail when you want. What was handed back is access to the process, not a claim on the calendar.

Here is the part headlines will skip, and the part that decides whether anyone loses money this quarter. The binding constraint is not the number of vessels in the queue. It is draft, which converts directly into payload, and thirty-three slots a day at a foot less draft is two separate taxes on the same transit. One you can buy. One you cannot.

Price the foot first, and put the assumptions in daylight. Assume a loaded Panamax bulker with a tonnes-per-centimetre immersion of about 60 tonnes at operating draft; substitute your vessel's hydrostatic table, every ship differs. One foot is 30.48 cm. Sixty tonnes times 30.48 is 1,829 tonnes of deadweight not available. Assume a 60,000 tonne grain stem and a voyage cost of $900,000 whatever the load. Full stem gives $15.00 per tonne. A 58,171 tonne stem gives $15.47 per tonne. One foot of draft costs $0.47 per tonne of freight, 3.1% on the unit cost, on a stem where the charterer was probably pricing to a margin thinner than that. In cargo value at $200 per tonne, the 1,829 tonnes left behind are $365,800 of product that does not move on this voyage.

Now the same sum in containers, because the arithmetic bites differently there. Assume a Neopanamax container ship with a tonnes-per-centimetre immersion of 110 tonnes, giving 3,353 tonnes for the missing foot. At 14 tonnes all-in per TEU, that is 240 TEU of nominal capacity stranded per affected transit. That is the headline number for the box trades, and it explains why forty-niner discounts matter more to liners than to bulk operators.

But beware of paying that cost before it is due. Draft binds only if you were going to load to fifty feet anyway. A box ship that floats at 47 ft on her stow plan loses nothing from a 49 ft limit: she was never going to use the foot. So the correct order of operations at every booking is your loaded summer draft against the permitted draft, and if the answer has margin, the foot costs zero and you can stop worrying about it. The 3,353 tonne figure is a worst case, not an invoice. Check the stow plan this week, before assuming damage.

Buy the slot, or run the risk? That is the second calculation, and it should be expected value, not hope. Assume a Panamax at $18,000 a day in time-charter equivalent terms, and take the advisory's own warning seriously enough to put an 80% probability on a material wait for an unbooked vessel. Expected delay cost equals probability times days times daily cost. Three days costs you $43,200. Five days costs $72,000. Ten days costs $144,000. Twenty days costs $288,000. The booking or slot price at which you stop bargaining is exactly that line: if the canal or the resale market wants less than your expected waiting number, pay it, and pay it the same day. Anything above that line you refuse, and you re-route instead. Put your own eighty-per-cent figure in and this becomes an arithmetic, not a negotiation.

For liners and their cargo the daily cost has a different shape, because hire is not the dominant term. Assume a 4,500 TEU Neopanamax sailing with $40,000 of cargo value per TEU, carrying cost at 8% annual. That is $8.77 per TEU per day, or $39,452 per day across the ship. Add $35,000 a day of charter cost. An expected wait of ten days at 80% probability costs carrying of $315,616 plus hire of $280,000, which is $595,616, or $132.36 per TEU. That is the number to compare against the alternative.

Assume a diversion or an alternative routing adds eight days of steaming; at $8.77 per TEU per day that is $70.14 per TEU of carrying plus additional bunkers. Under those two assumptions the certain routing costs less per TEU than the expected wait, by about $62. Run it with your own numbers, but run it: the answer is not always the same, and it flips as soon as your expected wait drops below six days or your cargo value per TEU drops below $21,000.

Nothing in the advisory prices certainty, so price it yourself. A regular booked slot is not a toll, it is an option on a date, and its value is the difference between a known transit and the distribution you live with without one. Back to the Panamax example: a booked slot converts an 80% chance of a five-day wait into no wait, which means it is buying back $72,000 of expected loss for whatever the fee happens to be. Nobody publishes that spread, which is why operators bid by feel, and why they either overpay wildly or walk away at a price they should have taken. Convert your own operation into an expected-waiting-cost figure once, write it at the top of the fixture sheet, and the slot debate stops being a mood.

There is a second owner of this problem who never touches the booking screen. A regular slot is a lead-time product, and its real cost shows up as inventory. Standard planning maths: safety stock equals service factor times lead-time standard deviation times daily demand. Assume a receiver drawing three TEU a day at $40,000 of value per TEU, aiming at a 95% service level, so a factor of 1.65. When canal transit was predictable, take lead-time standard deviation as one day, giving 4.95 TEU of safety stock. Push that deviation out to six days with an unbooked queue and safety stock goes to 29.7 TEU. The difference is 24.75 TEU, which is $990,000 of working capital sitting on a shelf earning nothing, bought to insure against a spread the canal just widened. Somebody funds that, and it is usually the party who never sees the reservation window. Put this figure in front of whoever decides to save the booking fee.

Grace Yang would skip the lock and go straight to what this does to the buyer's landed cost, so do that sum too. Take the grain stem. Freight rises from $15.00 to $15.47 a tonne, the stem is 58,171 tonnes, and the buyer pays $27,344 more in freight for a cargo 1,829 tonnes smaller. Spread across the delivered cargo that is a unit cost increase on the order of half a dollar a tonne, which at a $200 landed price is roughly a quarter of one percent. Manageable alone. Then add the canal's own 16.1% revenue-per-ton increase from fiscal 2026, which sits somewhere upstream in the toll structure, and the slippage in arrival reliability, which shows up not as freight but as demurrage, plant scheduling and safety stock. The buyer never sees a line called canal draft risk. He sees a delivery window that got wider, and a supplier asking for a force majeure clause.

Timing is short for some players and long for others, and the distinction matters. For a vessel already on the water or loading this week, the advisory is effective immediately: book or accept an unquantified wait, decided now. For bulk stems loading in the next three to six weeks aimed at Asian delivery, the decision has to be made at fixture, since the charter party decides who carries canal risk and a badly drafted clause puts it on whoever did not price it. For liner operators, tomorrow's schedule was fixed weeks ago and the damage this month is incidental; the exposure sits in the Q1 networks being finalised now, and in whether those strings assume 33 slots or something closer to normal. For anyone negotiating a 2027 contract of affreightment, this is the window: add a canal transit clause with a defined slot-booking obligation and a defined consequence, because the road implied force majeure arguments are where this ends up.

Who carries the cost differs by layer, and again it is uneven. Chartered-in tonnage pays the daily cost, voyage charterers pay it through time actually lost. Cargo owners under CFR or CIF terms find the cost inside the freight already quoted; under FOB they encounter it as a delay they did not buy. Just-in-time receivers feel it last and worst, because nothing upstream gave them a revised arrival date — the advisory specifically declines to guarantee one for unbooked vessels, and that refusal travels down the supply chain as absence of information rather than as bad news.

One caution about reading 521 million tons as recovery. A tonnage record tells you what happened to the average loaded vessel, not what happens to the next one trying to book. Allocation systems clear at the margin, and the margin here is thirty-three slots a day against demand that averaged 38.2 movements a day across the year. Averages describe what already happened; queues are built out of the tail.

Ask who benefits from scarcity, because this is not rhetorical and the published numbers answer it. Revenue up 16.6% on tonnage up 0.42% means the growth came entirely from unit revenue. A management team that has just delivered that result has a live interest in the conditions that produced it. No bad faith required: sell thirty-three slots into demand of thirty-eight a day, let the price find itself, and keep the queue count unpublished. The test is monthly and takes one minute: if revenue keeps rising while slots stay at thirty-three and draft stays at forty-nine feet, this has stopped being drought management and has become a product.

Exposure also needs naming lane by lane, because averages hide who is standing in it. The parties paying this are US Gulf grain and LPG stems going to Asia, container strings running Asia to the US East Coast, South American mineral cargoes moving north, and every regional service that built its rotation around a Panamax transit assumed to be bookable at short notice. For a pre-booked liner string the rest of this month is uneventful. For a spot bulk fixture worked out three weeks ahead, thirty-three slots a day at forty-nine feet is the whole deal, and the clause that decides who pays for waiting gets signed this week.

What would turn this around? Three scenarios, each with a testable tell. A wet November and December that breaks the 1.29 amplification, visible first in Gatún lake level rather than in rainfall totals; if the lake recovers, slots go up before draft does, and draft is what pays. A demand soft patch in the bulk trades, which would collapse the delay distribution even with 33 slots unchanged, meaning your booking premium should be recomputed monthly rather than treated as a constant. Or the canal taking more of its 16.6% revenue growth through tolls rather than slots, which converts a queue problem into a price problem and changes the arithmetic for cargo that is not in a hurry — for those shippers paying becomes rational again.

What gets decided this week is shorter than the analysis. Every agent and operator with November or December exposure should take three readings before the next fixture: current Gatún level against the advisory assumption, permitted draft on the booked date rather than today's, and actual queue length for unbooked ships in each lockage class. Queue length is the one number nobody publishes and every agent can obtain from the canal's own booking system or from the last two vessels through.

Then set a numeric rule and write it into the instructions given to whoever books. Mine, under my own assumptions, is this: pay any slot price up to $72,000 for a Panamax when the expected wait is five days at eighty per cent probability, and re-route rather than queue above that. The rule for liners reads much the same way: once expected delay cost per TEU exceeds the carrying cost of an eight-day longer routing, stop bidding and divert. Check your stow plan against 49 ft before assuming any payload loss at all, since the fleet ships that were already floating light lose exactly nothing.

For contract people, this is the cheap week to act. Every 2027 fixture should carry an explicit canal transit clause: who books, who pays the slot premium, what happens to laycan if the booking is not honoured, and whether delay caused by absence of a reservation counts as an exception. Waiting until there is a queue costs more than the clause does, and unlike the foot of draft, the clause is free.

Restoring last-minute bookings is reopening the last checkout lane in a shop that has sold out of stock: the queue gets shorter, the shelf does not. Thirty-three slots a day at forty-nine feet is a rationing plan with a polite face. The moment to lock was when the reservation window opened for your date, and the second-best moment is the day your expected wait number crosses $72,000.

↑ Back to the key points

— By Vivian Zhao