China-Europe airfreight rates firm as BAI30 gains 2% while Shanghai outbound slips 3.3%

The Analysis

I have flown cargo out of Hong Kong more times than I can count, and the thing I always tell people is that an airport is not a point on a map, it is a road with a runway at the end of it. The week ending 24 August only makes sense if you look at it that way. TAC Index put the Hong Kong outbound number, BAI30, up 2% on the week and 22.8% above where it sat a year ago, while Shanghai outbound, BAI80, came off 3.3% on the week and stayed 18.5% higher year on year. Two cities, one week, two different roads. If your cargo is in the Pearl River Delta you are driving the Hong Kong road. If it is in the Yangtze Delta you are driving the Shanghai road, and this week those two roads pointed in slightly different directions.

The market underneath those two numbers is worth a minute of your time. The European Union put a temporary flat duty of 3 euros on low-value parcels on 1 July, and the weeks right after were ugly for anyone whose business was built on small parcels going straight from the mainland or Hong Kong into Europe. Volumes dropped and spot rates dropped with them. What the week to 24 August shows is that the drop has stopped. China-Europe rates firmed while China-US rates went soft, which tells you the adjustment is largely done and the lane has found a new shape. Bangkok and Hanoi into Europe firmed as well. India was roughly flat on the week but sharply higher than a year earlier. None of that is what a market in retreat looks like.

The global Baltic Air Freight Index, BAI00, slipped 0.2% on the week and is still 21% above last year. That is the number I watch when I want to know whether capacity is genuinely loose anywhere, and the answer is that it is not. The situation in the Middle East, the extra distance aircraft are flying around Hormuz, and jet fuel that will not come down are all still sitting underneath the global figure, and they are the reason I do not believe anyone should be waiting for air space to get cheap before the fourth quarter. This is not a piece about predicting rates. It is about which road you put your boxes on this week, what the border crossing costs you in days, and how to tell when the detour has stopped being worth the fuel.

I want to start with the road, because that is where the money actually leaks. A shipment out of Hong Kong and a shipment out of Shanghai are not the same product even when they land at the same European airport. Hong Kong outbound covers spot and contract together, which is why BAI30 tends to move the way a negotiated book moves: slower on the way up, stickier on the way down. Shanghai outbound fell 3.3% in a single week while still sitting 18.5% above last year, and that is what a spot-heavier lane looks like when a few extra bellies and a couple of charters get absorbed into the market. If you have been treating these two origins as interchangeable lines in your routing table, this week is a good argument for splitting them into separate entries with separate owners.

Here is the practical version. Assume you are shipping 3,000 kilos a week of mixed consumer goods out of the Pearl River Delta, roughly half of it e-commerce parcels and half of it replenishment for a European warehouse. Your two realistic air paths are truck to Hong Kong and fly, or truck to Shanghai and fly. The truck to Hong Kong is a border crossing you can complete inside a day when the queue at the crossing behaves, and I have had weeks where it took six hours and weeks where it took twenty. The truck to Shanghai is about two days of road with no border, but you are adding road cost and adding a day of a different kind of uncertainty, the kind that comes from weather and from a driver who has to stop.

Now put the flying on top. Assume a Hong Kong departure gets you wheels-up within a day of the cargo being ready, and a Shanghai departure needs two days to the runway because of the road leg. Once it is in the air, both are roughly the same to a European gateway, and the difference at the far end is a day of handling one way or the other. So your honest door-to-gateway estimate is about four to six days through Hong Kong and five to seven days through Shanghai, and that one-day spread is the whole argument. On a 3,000 kilo book, one day is not a rounding error if your European warehouse is running on a seven-day reorder cycle.

The other thing that changed the shape of this lane is the European Union's temporary flat duty of 3 euros on low-value parcels, in place since 1 July. I am not going to walk through how it is assessed or who bears it, because that is not my job and your compliance people should be reading the actual text. What I care about is that it changed the piece count economics of the lane, and the weeks immediately after 1 July were exactly what you would expect: volumes out of the mainland and Hong Kong into Europe fell, and spot rates fell alongside them. The market spent six or seven weeks digesting that. The week to 24 August says the digesting is finished.

Let me do the arithmetic the way I would do it on a whiteboard with a customer. Assume one of your weekly parcels is 400 individual low-value packets going to four European countries. If each of those packets is the kind that picks up the flat 3 euro charge, that is 1,200 euros sitting on top of the freight before you have paid for a single kilo of lift. Assume those 400 packets weigh 800 kilos all in. If you consolidate them into, say, eight declared shipments into a single European gateway and break them there, the piece count that the flat charge attaches to disappears from your calculation, and you have traded 1,200 euros of per-piece exposure for the cost of eight declarations plus a break-bulk operation at the far end. Whether that trade is good depends on two numbers you can get hold of this week: what your European break-bulk partner charges per shipment, and how many days the break-bulk adds.

Assume the break-bulk adds three days at the European end. Three days on 800 kilos of consumer goods is not nothing, but it is not 1,200 euros either unless your goods are extremely valuable per kilo. At a declared value of 25 euros a kilo, 800 kilos is 20,000 euros of goods, and three days of carrying that at an 8% annual cost of money is roughly 13 euros. That is the kind of arithmetic that makes consolidation look obvious. But if your goods are worth 200 euros a kilo, the same three days costs you about 105 euros, and you also have to ask what three extra days does to a customer who was promised a four-day delivery. The point is not that consolidation always wins. The point is that a flat per-piece charge turned a freight question into a piece count question, and piece count is something your operations team controls, while your freight rate is not.

So where does that leave the Hong Kong versus Shanghai decision for the rest of this month? The week-on-week gap between BAI30 up 2% and BAI80 down 3.3% is about five points in seven days. That is real, but it is one week. Both indices are still dramatically above last year, 22.8% and 18.5% respectively, and that is the number that should be driving your budget rather than the weekly wobble. If your routing rule is to chase whichever origin printed lower last week, you will spend the next eight weeks moving cargo back and forth between two provinces and losing more in road cost and extra handling than you save in air.

Here is the rule I would write instead. Split the book by geography and let it stay split. Everything originating south of the Nanling goes to Hong Kong. Everything originating in the Yangtze Delta goes to Shanghai. Then set a single override: if the weekly gap between the two indices stays above four points in the same direction for two consecutive weeks, move no more than 20% of the book to the cheaper origin and hold it there for at least three weeks before you review. That is boring, and boring is what you want in August. The reason to cap the move at 20% is that a big origin switch creates its own congestion. Your Hong Kong trucking partner suddenly needs twice the trailers, your Shanghai forwarder suddenly needs twice the pallet positions, and both of them will charge you a premium for the privilege of being surprised.

Let me talk about the border for a minute, because people underestimate it every single year. The Hong Kong land crossings are not a formality. On a normal weekday the queue is a known quantity and your driver plans around it. In late August you have two things stacked on top of each other, the tail of the summer peak and the start of the Q4 pull-forward, and on top of that you have typhoon season, which in the Pearl River Delta can close a crossing or a terminal with about twelve hours of warning. I have sat in a truck park in Shenzhen waiting for a crossing to reopen and watched an entire week of bookings collapse into one very expensive Friday. If your cargo is routed through Hong Kong over the next six weeks, you need a written contingency that says what happens when the crossing closes, and it needs to name an alternative rather than saying that someone will advise.

The alternative worth naming is the road north. If the parcels and the small consolidated shipments can tolerate eighteen to twenty-two days instead of five, the rail corridor through Alashankou and Khorgos into Malaszewicze and on toward Duisburg is a completely different road with completely different risks. I am not going to quote you a rate for it, and nobody should quote you a transit time on that corridor as a promise either, because its variability lives in the gauge change at the border and in the European leg, not in the Chinese departure. But the arithmetic is simple enough to be useful. Assume your air path is five days door to gateway and the rail path is twenty days, and remember that twenty is my planning assumption and not anybody's commitment. That is fifteen extra days of inventory in transit.

Fifteen extra days of inventory in transit is a real cost, and it is the cost people forget to write down. Assume your goods are worth 40 euros a kilo and you are moving 3,000 kilos a week, which is 120,000 euros of goods. At an 8% annual cost of money, holding that fifteen extra days costs you about 395 euros a week, and across a quarter that is roughly 5,100 euros. Against it you save the entire air freight bill on that volume. Whether the trade is good depends on your air bill, which I am not going to guess at, but the structure of the decision never changes: rail buys you a lower freight line and charges you in transit inventory and in variability, while air buys you predictability and charges you in freight.

If your European warehouse is running lean, air is still the cheap option even at 22.8% above last year. If your European warehouse has room and your demand forecast is stable, the rail corridor starts to look sensible for the replenishment half of the book and not for the parcel half.

That split is the thing I would actually push. Take the replenishment volume, the goods going into a warehouse with no customer waiting on a specific delivery date, and test it on the rail corridor for one full quarter. Take the parcel volume, the goods with a consumer on the other end and a delivery promise attached, and keep it on air while you fix the piece count problem instead. Trying to solve both with one mode is how companies end up with a warehouse full of goods nobody ordered and a customer base that stopped believing the delivery estimate.

Now for the reason I want people acting this week rather than next month. The global index is 21% above where it was a year ago and it barely moved on the week, down 0.2%. That combination, high year on year and flat week on week, is not a market about to hand you a discount. Behind it sit the Middle East situation, the longer routings around Hormuz, and jet fuel prices that have not come down. Every one of those shows up in your surcharge line rather than in your base rate, which is exactly how an air freight bill quietly gets more expensive while the index itself does very little. If you are waiting for a better week to book fourth quarter space, assume the better week is not coming and price the quarter on today's numbers.

On surcharges, here is the pitfall I see most often. A forwarder quotes you a rate and the quote carries a fuel and security line described as 'as published' or 'at current levels'. That line is not fixed. It moves with fuel and it moves with routing, and on lanes being flown longer because of the Hormuz situation it has been moving in one direction. When you compare two quotes this week, ask for the surcharge to be stated as a number with a validity date, and ask what happens to it if the routing changes. A quote that is 4% cheaper on the base line and 8% more exposed on the surcharge line is not a cheaper quote, and in a quarter like this one it is usually the more expensive one. Get it in writing with a date on it.

The other pitfall is the transit promise itself. The five days door to gateway I used above is an estimate with assumptions attached, and I labelled it that way on purpose. When a carrier or a forwarder gives you a number, ask what the number is measured between. Gateway to gateway, airport to airport, and door to door are three different figures, and the gap between them is usually the two days of road at the Chinese end plus a day of handling at the European end. I have watched a customer argue with a forwarder for two weeks about a late shipment when the real disagreement was that one of them was counting from the factory gate and the other from the aircraft. Write the two endpoints into the booking. It takes five minutes and it saves the argument.

At the European end, the gateway you pick matters more than most people think, and it matters for a routing reason rather than a price reason. If your break-bulk partner sits in the Netherlands and you are flying into Frankfurt, you have inserted a road leg into the middle of a process that was supposed to be fast. If your parcels are going into four countries, the gateway with the best onward trucking density is worth more than the gateway with the cheapest handling. Assume your onward European leg is two days from a well-connected gateway and four days from a cheap one. Two days on a parcel that was promised in four is a broken promise. Choose the gateway by the road that comes after it, not by the handling fee.

So here is what I would have done by the end of this week, and by this week I mean before Friday 4 September, because the data we are working from stops on 24 August and is already going stale. The routing manager should have the book split by origin geography with a weekly kilo figure attached to each half, and that split should be written down rather than living in somebody's head. The pricing lead should go back to the top two forwarders and ask for surcharge lines restated as fixed numbers with a validity date, plus a written statement of which two endpoints the transit estimate is measured between. The operations lead should get the Hong Kong crossing contingency written down with a named alternative, and should confirm with the trucking partner how many extra trailers they can actually produce on 48 hours notice, because the answer is usually fewer than you assume.

In the week of 7 September, the same three people should run the piece count exercise on the parcel half of the book. Take one representative week, count the pieces, multiply by 3, and put that figure next to the cost of consolidating into declared shipments and breaking bulk in Europe. If the flat duty line comes to more than 10% of your total European parcel cost, consolidation is worth testing on 20% of parcel volume for four weeks. If it comes to less than 5%, leave the channel alone and spend the effort on the delivery promise instead, because that is where your customer actually feels the difference. By 21 September the rail test should be live for the replenishment half, with a booked departure and a measured transit rather than a quotation.

The thresholds I would write into the routing rule are these. If either BAI30 or BAI80 prints a weekly move larger than 4% in either direction, freeze origin switching for two weeks and simply observe, because a move of that size in a single week is usually a data artefact or a one-off charter and chasing it will cost you more than waiting. If the year-on-year gap on either index closes to below 10%, that is the signal the lane has genuinely loosened and you can start negotiating contract space again with some leverage. Until that happens, assume you are buying in a tight market and buy earlier rather than later. And if your forwarder's Hong Kong trucking partner misses the crossing window twice inside a week, that is not bad luck, that is a capacity problem, and the phone call you need to make is about trailers, not about bookings.

I have said this before and I will keep saying it: most of the money on this lane is not lost on the aircraft, it is lost on the road at either end of it. The week to 24 August gave us two airports moving in slightly different directions, a parcel duty that has finished being digested, and a global index sitting 21% above last year and going nowhere. None of that is a reason to panic and none of it is a reason to wait. It is a reason to write down which road your cargo is on, what the border crossing costs you in days, and what you do when it closes. Change the tire before you need it, not on the hard shoulder at two in the morning with a truck full of parcels sitting behind you.

↑ Back to the key points

— By Ömer Kaya