Baltic air index slips 2.1% to Oct 12 but stays 24.3% up YoY as jet fuel doubles

The Analysis

Open with the map and not with the chart. The TAC Baltic air cargo index for the week ending 12 October slipped 2.1% against the week before and still stands 24.3% above the same week a year earlier. Those two numbers are not arguing with each other. One of them is a week long, the other is a year long, and reading either without the other is how a shipper ends up signing a 2027 rate in the wrong month on the wrong assumption.

I spent the front half of a working life selling and operating freighter space out of Pudong and Hong Kong, and the back half moving the same categories of cargo on China-Europe block trains and on trucks that queue at Khorgos and Alashankou. That is why this column never reads a rate print as a market call. The question is which node on the line moved, because in air freight the node is the market: one gateway out, one gateway in, one customs house between them, one truck fleet at the far end. Everything else is commentary.

So the week to 12 October gets read here the way a dispatcher reads it. Shanghai outbound off 2.1%. Hong Kong up 0.2%. Frankfurt up 3%. London Heathrow up 12.1%. Chicago outbound to Asia down 29%. Jet fuel up 111.2% year on year after another 3.1% in the week to 9 October. Five nodes moving in five directions is not noise. It is a routing problem, which is to say it is a money problem, and routing problems have answers that a headline does not.

Start with the arithmetic nobody prints. The 2.3 points between Shanghai and Hong Kong, the 9.1 points between Frankfurt and Heathrow, and the 29 points under Chicago are the actual story. The 2.1% headline is an average of them, and an average is where useful detail goes to die. A Baltic print is a basket; your shipment is not a basket. Your cargo leaves one gateway, lands at one gateway, clears at one customs house, and is priced off that pair of nodes and nothing else.

The weekly fall is the calendar talking. Golden Week shuts the factories, and freight does not stop on the same day the lines stop: the last push goes out in the final days of September, and the week to 12 October is when the hangover lands in the booking system. Space that was scarce on 28 September is briefly unsold on 8 October, and a broker index picks that up within days. That is a seven-day event and it will be gone in seven days. The 24.3% year-on-year gain is the number that carries a signature, because a gap of that size is not a holiday. It is a structure, and structures are what you write contracts against.

Vivian Zhao reads a print like this from the rate desk rather than from the map, and she would go straight to the money: with the index sitting 24.3% above last year, any annual rate you sign for 2027 off the current spot level is being signed on a market that is a quarter above last year, not on a market that fell 2.1% last week. The weekly number is the one that will be quoted back at you in a negotiation. The annual number is the one that will show up on your invoice.

Before any of that, settle who is actually carrying the number, because on an index that sits 24.3% above last year the answer decides whether you have a problem this month or next year. If you buy ex-works or FCA out of the delta and you control the air leg yourself, the 24.3% is on your account now and the 2.1% is the whole of your relief. If you buy on delivered terms to a European door, the increase is sitting on your supplier's or your forwarder's account today, and it will reach you at renewal, with something added for the trouble, because that party has been funding it since September. Importers who believe delivered terms insulate them are usually only deferred, and the deferral ends on the day the contract comes up for review.

Take the cheapest node switch on the board first. Shanghai outbound is off 2.1% and Hong Kong's overall index is up 0.2%, a 2.3-point gap between two gateways that serve the same manufacturing hinterland and sit inside overlapping trucking and barge radii. Do not underestimate that gateway. Hong Kong runs its own freighter fleet with its own rotation pattern, its own belly capacity on long-haul passenger services, and a customs regime that behaves differently on the same commodity lines. Moving a shipment from Pudong to Hong Kong buys you 2.3 points on paper and costs you an inland leg, a second handling, a broker who may not know your codes, and a day or two that you may not have.

The trade only works if the inland leg is already in your system. If you are shipping out of Suzhou, Ningbo or the Pearl River delta and you already have a bonded truck or a barge feed into Hong Kong, the 2.3 points is real money for no operational change. If you have to build the inland leg from scratch to collect it, you are paying in days and in exception risk to save two points on a lane, and that is the sort of arithmetic that looks clever in a spreadsheet and ugly in a month-end review.

One caution on that 2.3 points, because it is the sort of detail that gets quoted wrong for a whole season. The print gives Shanghai outbound and Hong Kong's overall index, and those are not the same cut of data: one is a directional reading out of a single gateway, the other is a composite across that gateway's trade. Do not book the gap as though it were measured on the same lane. Take it as a direction and then price your own pair of nodes, Shanghai to Frankfurt against Hong Kong to Frankfurt, with your own forwarder, at your own weight break, on your own commodity.

Now the European end, where the gap is four times as wide. Frankfurt is up 3% on the week and Heathrow is up 12.1%, which is 9.1 points between two airports that are, for most European distribution, one truck day apart. Put plainly, this is a road question. If your UK and continental volumes can be landed at Frankfurt and trucked across, the 9.1 points is available to you and the price of it is a day of road transit plus a customs entry on the UK side. If your goods must clear in the UK because of the way your buyer's entry is set up, the 12.1% is simply your number and there is no node to switch to.

Run that on a real figure. Assume you move eight tonnes of general cargo a month into Europe and your September air bill on the Heathrow routing is 100. The same weight landed at Frankfurt and trucked in saves you the 9.1-point gap on the air leg and costs you the road leg, which on eight tonnes is a single truck movement and an entry. If the road leg plus the extra handling costs less than 9.1% of your air bill, the switch pays. If your trucker is quoting you peak-season spot rates for that crossing because everyone else had the same idea in the same week, it does not. The tell is whether the road quote moved in October; if it did, the congestion has already followed the arbitrage.

Chicago is the one to stare at. A 29% fall on Chicago outbound to Asia is not a rounding error and it does not happen on a lane with balanced demand. It says the transpacific has been running full eastbound and thin westbound, and that US outbound capacity into Asia is being sold at whatever it will fetch. If you move anything back the other way, repair parts, returns, samples, re-export of returned goods, tooling going back to a contract manufacturer, this is the week to book it and to book it for the whole fourth quarter if a carrier will let you.

Do not build a plan on it, though. A lane that falls 29% is a lane with surplus capacity, and surplus capacity is the first thing a carrier pulls when the peak fills in. Treat the 29% as a spot opportunity with a hard expiry, take what you can move in the next two to three weeks, and do not renegotiate an annual programme on the assumption that Chicago outbound to Asia will still be 29% down in February.

Peak never arrives as a rate. It arrives as a booking that will not confirm. On the lanes this column used to operate, the first symptom is not the price, it is the cut-off: the cargo terminal stops accepting at the published time because the dock is full, the trucker cannot get a slot, the build-up of the pallet slips by a shift, and the aircraft leaves with your freight still on the ground. Rolling is how a market that looks 2.1% softer ends up costing you ten days. For the next three weeks, watch the roll rate on your own bookings rather than the index; two rolls inside one week on the same lane means the node has closed, and it closes well before the price admits it.

The fuel line is where the story turns, and it is the line most readers skip. IATA jet fuel is up 111.2% year on year and put on another 3.1% in the week to 9 October alone. Fuel does not enter your bill through the base rate. It enters through a surcharge that runs on its own schedule, on its own formula, and in most carrier tariffs on its own revision calendar, weekly or monthly. That is the mechanism by which a carrier whose fuel bill has doubled can watch its headline rate fall 2.1% and still be collecting more money from you than it did last week.

Here is the thing the reports did not print, and it is the one that actually costs money: the Baltic index is not your invoice. It is a spot broker rate. Fuel surcharge, security, war risk and peak-season surcharge sit on top of it and are billed separately, and the single largest of those line items is now moving on a fuel price that has more than doubled year on year. So the direction of the index and the direction of your landed freight cost have come apart, and they have come apart by more than the index moves in an ordinary week.

Put it on the eight-tonne example. Assume your September invoice, all line items in, is 100, split 80 base rate and 20 fuel surcharge, and assume your surcharge tracks jet fuel. The index leg goes up 24.3% year on year, so 80 becomes 99.4. The surcharge leg goes up with fuel at 111.2%, so 20 becomes 42.2. Your invoice is 141.7, while the index is telling a story of 124.3. That is 17.4 points, roughly 12% of what you actually pay, sitting outside the number everybody is quoting. Apply the 2.1% weekly fall to the base rate and you claw back about 2.1 points. You gained two and the fuel took twenty-two.

If you buy capacity the way this column used to sell it, on a block space agreement or a charter, the fuel clause is the entire ball game. A BSA with a fixed fuel surcharge is a gift in a year when fuel has doubled. A BSA with a surcharge that floats on an index is a blank cheque, and it is the clause that decides whether your annual deal is worth the paper. Ask your carrier one question before you sign: what is the reference, what is the revision frequency, and is there a cap. If the answer is a weekly floating reference with no ceiling, you have not bought a rate, you have bought a subscription to a fuel price you do not control.

The sentence in the TAC note that deserves more attention than the index itself is the one about AI hyperscalers whose space is not yet committed. Anyone who has sat on the sales side of a freighter knows what an uncommitted block of that size means. Hyperscaler cargo does not arrive as loose boxes looking for a spot rate. It arrives as a programme: whole pallet positions, whole main decks, often whole rotations, booked on annual agreements and taken out of the market in one signature. When that signature lands, it does not nudge the index by two points. It removes the capacity that everyone else was planning to bid on.

That converts the rest of October into a question of timing rather than price. Capacity that looks available on 20 October because the hyperscaler space is still open can be gone by mid-November. If your fourth-quarter peak plan assumes you will be able to buy space in the second half of November at something close to today's level, you are making a bet on someone else's procurement calendar. This is the part of the peak that nobody can price for you, because the only people who know the date are the ones placing the order.

Which brings the discussion back to the question that always closes a routing argument: does this lane still walk? Air freight out of the Yangtze delta into Europe is a five-to-seven-day door-to-door proposition in normal weeks and a fifteen-to-twenty-day one when the peak closes and your cargo gets rolled. Rail on the same corridor is measured in weeks rather than days and costs a fraction per kilogram, and the truck corridor through the Khorgos and Alashankou crossings sits between them. The substitution is not free and it is not instant, but it is the only real hedge you have against a capacity squeeze that is being driven by somebody else's booking.

Set the substitution up before you need it. Split your fourth-quarter volume so that a defined share, say the non-urgent 20% by weight, has a rail or truck routing that has already been tested once this year with your own cargo and your own documents, not as a quotation but as a shipment. A routing you have never run is not a hedge, it is a wish. The cost of running the test is one shipment; the cost of discovering in week 48 that your rail routing fails on a document you have never filed is a season.

Then put triggers on it rather than dates, because a date assumes you can see the future and a trigger does not. Open the rail or truck option when your air rate on the Shanghai to Europe lane moves another 5 points above today's level, or when your forwarder rolls you twice in one week on the same booking. Close it and go back to air when the roll rate drops back and the rate comes off. Write the trigger into your logistics plan now, while the market is quiet enough to think, because the week you need it is the week nobody has time to design one.

And watch the fuel line more closely than the rate line for the rest of the year. A 3.1% move in jet fuel in a single week, on top of 111.2% year on year, is the number that will hit your budget first and hardest. It is also the number that is easiest to negotiate against if you know it is coming, because surcharges are published in advance and can be planned around, while spot rates cannot. Ask your forwarder for the surcharge calendar in writing and build it into your fourth-quarter landed cost model rather than reading it off a bill after the fact.

The line still walks. It has walked through every peak I have loaded a freighter for, and it will walk through this one. What has changed is the price of standing on it without a plan. Shanghai is two points cheaper than it was last week and a quarter more expensive than it was last year; Frankfurt is nine points cheaper than Heathrow if your paperwork lets you land there; Chicago to Asia is 29% off if you can use it this month and not a basis for anything if you cannot. Read the nodes, not the average, and the week to 12 October stops being a confusing headline and becomes what it really is: a set of choices with prices attached.

↑ Back to the key points

— By Ömer Kaya