YTO runs second Budapest-Jiaxing bonded transfer on 29 August after 28.3-tonne pilot

The Analysis

On 26 August a YTO Cargo flight out of Budapest put 28.3 tonnes down at Jiaxing Nanhu Airport, and the interesting part was not the landing. It was what happened next. Under Jiaxing Customs supervision the cargo went onto a bonded truck and ran to Shanghai Pudong, the first international air leg plus supervised road import transit through YTO's Oriental Worldport hub. Two more Budapest to Jiaxing import transits were put on the board for 27 and 29 August. For a Yangtze Delta importer this is not a story about an airline. It is a second answer to the question of where a file gets cleared, and a second answer is worth real money when the first answer is Pudong in a peak week.

I have spent enough time on transit files to know what people get wrong about them. They think transit is a trucking arrangement with some paperwork attached. It is the other way round. Transit is a customs procedure that happens to involve a truck, and the truck is the least of your problems. What matters is that the arrival point and the clearance point are no longer the same place, which means two customs offices touch the same file, on two clocks, with two sets of records, and the thing holding them together is a seal number and a document trail. When those two agree you have gained a day and a half. When they do not you have gained a stopped truck.

The tonnage is small and that is fine. Twenty-eight point three tonnes on one rotation is a pilot, not a service. The hub opened its first international cargo route in May and put a regular international truck schedule on the road in July, so this is a hub building itself in public, one capability at a time. I would plan against the capability I can actually see and keep the old path warm rather than the other way round.

Here is the file, and it is worth walking end to end because every handover is a place where the papers can stall.

A consignor in Central Europe books 1,800 kilos, eight pallets, onto YG9026. The commercial set, meaning invoice, packing list, contract and any certificate of origin the goods need, is produced by the shipper abroad and is the only place the value and the origin actually live. The master air waybill and the manifest come from the airline's cargo side or its agent, and they are what tells the arrival airport what is on the aircraft. The importer's side produces the declaration package: the classification, the relief treatment, the consignee's identity, the authorisation to file, and any licence or certificate the commodity needs. The transit itself then adds a fourth document, the transit declaration, plus the record of the vehicle and the seal that moved the goods.

Four sets of papers, four parties, and the failure mode is always the same shape: three of them agree and one does not. A master that says 28.3 tonnes against a set of house documents that add up to 27.9 tonnes is a problem you cannot fix from your side, because the master belongs to the airline and the cut-off for changing it usually sits before departure. A packing list that says eight pallets and a transit declaration that says nine is a problem you also cannot fix quickly, because by then the goods are sealed on a truck and the seal is a matter of record at both ends.

So the first operational rule on a transit file is that reconciliation happens before the aircraft lands, not after. Get the master and the house documents to agree on piece count, gross weight and description while the goods are still in the air, because the window in which a manifest change is cheap closes on arrival, and the window in which it is merely expensive closes when the seal goes on. If you take one sentence out of this piece, take that one.

Classification comes before relief here too, the tariff number first and the relief second. I have watched a transit file lose two days because somebody decided the goods qualified for a particular treatment, built the declaration around it, and only then checked whether the tariff number underneath supported it. It did not, and by then the goods were already at the destination supervised site, which is the worst possible place to discover a classification error, because amending there means reopening a record that has already been closed at the other end. Argue about the number while the aircraft is still in the air. Argue about everything else afterwards.

Now the clock. Assume one consignment of 1,800 kilos, eight pallets, arriving at Jiaxing on YG9026 with wheels down at 06:00, then moving under transit to a supervised site at Pudong. Assume ground handling and manifest confirmation take three hours. Assume the transit declaration is accepted in one hour. Assume supervised loading and sealing take one hour. Assume the road leg from Jiaxing to Pudong takes two and a half hours. Assume handover and seal verification at the Pudong end take an hour. Assume the destination declaration is filed in two hours. Assume release, with no exam, takes three hours. That is thirteen and a half hours from wheels down to release, and it is a same-day file if nothing goes wrong.

Then the exam case, because that is the one that decides whether the model is worth using. Assume an exam rate of eight per cent and assume an exam adds one and a half days. Expected delay across the book is about 0.12 days a shipment, which sounds trivial until you remember it lands on the shipments that matter, the ones with a production line waiting. On those, one and a half days is the whole conversation, and no amount of average performance will console the plant manager whose line stopped.

Put money on it, with the assumptions visible. Assume a Yangtze Delta importer moving 40 tonnes a month out of Central Europe, in 22 shipments of about 1,800 kilos, historically clearing at Pudong. Assume Pudong dwell from arrival to release of three and a half days and three free storage days, so half a charged day. Assume storage at forty-five cents per kilo per day. Half a day on 1,800 kilos is four hundred and five dollars. Assume the exam rate above adds an expected 0.16 charged days, call it a hundred and thirty dollars. Pudong storage comes to about five hundred and thirty-five dollars a shipment. Assume the truck from Pudong to the plant in Suzhou costs two hundred and sixty dollars.

Run the same shipment through Jiaxing. Assume dwell before the transit declaration is accepted is 0.6 days and the supervised move adds another 0.2 days, so under a day in total, inside the free period at Jiaxing, so no storage charge on these assumptions. Assume the supervised vehicle from Jiaxing to a bonded site near Suzhou costs a hundred and eighty dollars. Assume the transit itself costs money as well: ninety dollars for the transit declaration and sixty for the vehicle supervision, a hundred and fifty all in. The Jiaxing route saves five hundred and thirty-five in storage and eighty in trucking, and costs a hundred and fifty in transit fees. Net four hundred and sixty-five dollars a shipment. At twenty-two shipments a month that is a little over ten thousand dollars a month, call it a hundred and twenty-three thousand a year.

That number is real and it is also fragile, and those two things are the same thing. It depends on one flight number. YG9026 is one rotation. The 27th and the 29th were scheduled, and nothing in what I have read commits anyone to a frequency beyond that. If the rotation drops, the transit capability does not degrade gracefully, it disappears, and the operating procedure you wrote around it becomes a document describing a service that no longer exists. So the rule I would apply is unglamorous: move the volume you can afford to lose, keep at least thirty per cent on the Pudong path, and reprice the whole thing every quarter against the frequency that actually operated rather than the one that was announced.

There is a second value here that is harder to put a number on and that I think matters more for most importers: the clearance point becomes a choice. If you can complete the transit formalities at Jiaxing and then move the goods on to any bonded site, you are no longer obliged to clear where the aircraft happens to land. That lets you match the clearance point to where the goods are actually going, which for a Yangtze Delta book means Suzhou, Wuxi, Hangzhou and Ningbo instead of one choke point at Pudong. Anyone who has watched a file sit at Pudong in a peak week knows the value of a second door is not the four hundred dollars. It is the option.

This one depends on the port, and I want to be careful about that. The dwell assumptions, the free period, the supervised vehicle fees and the waiting-time charges are set site by site, and the number at one supervised site is not the number at the next. Do not take my three and a half days and my forty-five cents as facts about your operation. Take them as the shape of a calculation and then go and get your own numbers in writing from the sites you actually use.

Which brings me to the alternatives, because a transit is not the only answer to a congested clearance point. Assume a shipment of 1,800 kilos with a goods value of twenty-five dollars a kilo, so forty-five thousand dollars. Assume air from Central Europe at four dollars twenty a kilo, seven thousand five hundred and sixty dollars, and three days origin to arrival. Assume rail at two dollars ten a kilo, three thousand seven hundred and eighty dollars, and eighteen days. Assume sea at ninety cents a kilo, one thousand six hundred and twenty dollars, and thirty-five days. Assume carrying cost on the goods at twenty per cent a year, about twenty-five dollars a day on forty-five thousand dollars.

Add the two together and the picture changes. Air is seven thousand five hundred and sixty plus nothing, so seven thousand five hundred and sixty. Rail is three thousand seven hundred and eighty plus fifteen extra days of carry at twenty-five dollars, three hundred and seventy-five, so four thousand one hundred and fifty-five. Sea is one thousand six hundred and twenty plus thirty-two extra days, eight hundred, so two thousand four hundred and twenty. Against air, rail saves about three thousand four hundred a shipment and sea saves about five thousand one hundred and forty.

The saving is not free and it is not available to everyone. It is available exactly to the extent that you know the demand in advance. Assume you can forecast sixty per cent of your volume six weeks out. On twenty-two shipments a month that is thirteen shipments that could move on the slower modes, at a blended saving of roughly four thousand three hundred each, about fifty-five thousand nine hundred a month. Assume the extra pipeline costs you six weeks of additional safety stock, say an extra hundred and eighty thousand dollars tied up, at twenty per cent a year, three thousand a month. The trade still lands heavily on the side of moving it. But the other forty per cent, the breakdown orders and the line-stop orders, will always be air, and those are precisely the shipments where a half-day saving at the clearance point is worth more than three thousand dollars of freight.

The fourth option is the one people forget: do not decide yet. If the goods can go into a bonded site and the clearance point can be chosen later, you have bought yourself the time to find out where the demand actually is before you commit the entry. That is worth something when your customers change their call-off dates, and it is worth nothing at all if your goods have to be on a line on Tuesday. Whether deferral of duty and taxes applies to your particular facts is a question for your filer and not for me, and I would ask it before I built a plan on it.

The pitfalls on this file are the ones that always bite on a transit, and I have met every one of them.

The seal is the first. If the seal number on the transit declaration does not match the seal on the truck, or the seal is broken in transit, the movement stops and you are into a procedure measured in days rather than hours. Photograph the seal at both ends. Record the number on your own file. Do not rely on the operator's record alone, because when there is a discrepancy the operator's record is the one being questioned.

The amendment clock is the second, and it is the one that costs real money. On a transit file there are three windows and they are not equal. Before the aircraft lands, changes are cheap and mostly yours to make. Between landing and acceptance of the transit declaration, changes are possible but both ends have to move. Once the goods have left the supervised site at the destination, you are no longer amending a transit, you are dealing with goods that have already been released, and that is a slower and more formal process with a different set of records behind it. Finish the classification argument before the aircraft lands. Everything you argue after that costs more.

The charging basis is the third. Free storage, supervised vehicle waiting time and any surcharge on the transit leg all have to be confirmed in writing for the specific site you use. Ask for three numbers: the free period in days, the rate per kilo per day after it, and the waiting-time charge for the vehicle if your paperwork is not ready when the truck is. Those three numbers are the difference between a routing that saves four hundred dollars and one that saves nothing, and they vary by site.

Who is allowed to touch the manifest is the fourth. On a consolidated master your house data is one line inside somebody else's document. If your weights change after the master is filed, the change has to go through the carrier's agent, and their cut-off is not your cut-off. Get that cut-off in writing and put it on your own booking instruction, because the airline will fix it is not a plan, and by the time you discover it was not fixed the goods are on a truck with a seal on them.

Responsibility and records are the part that decides what happens six months later, when somebody asks why a shipment was two days late or where a pallet went. Five things on every transit file: the master and house reference pair, the transit declaration reference with the acceptance timestamp, the seal number with photographs at both ends, every amendment with the reason and the hour it was made, and the name of the person who released the data on your side. Sample ten files a month and read them against the timestamps. If the gap between landing and transit acceptance is growing, you will see it there before you see it in a customer complaint.

And there is a boundary I am not going to cross. Which goods can move under a supervised transit at all, what the conditions are, and where the exemptions sit are questions of rule, not of routing. The applicability and the exceptions on this kind of supervised movement are Dr. Ingrid Voss's ground, not mine. What I can tell you is where the file sits, who holds it at each handover, and how long each step takes when it goes well, and that is the part I will own.

The inventory consequence is smaller than people expect and I would rather say so than oversell it. Assume your pipeline from Central Europe to the plant is forty-two days door to door on the air and Pudong path, and assume the transit path takes one day out of it, which is what the thirteen and a half hour file suggests on a good day. Assume consumption of forty tonnes a month at twenty-five dollars a kilo, so a million dollars a month. Assume you hold three weeks of safety stock against demand variability, seven hundred and fifty thousand dollars.

A day out of the pipeline lets you take roughly a day of consumption out of the buffer, about thirty-three thousand dollars, and at twenty per cent carrying cost that is worth around six thousand seven hundred a year, five hundred and fifty-six a month. Set that against the ten thousand a month of fees and storage the transit saves and the ratio is clear: this is a document and dwell play, not an inventory play. Anyone who sells it to you as a working capital story is pointing at the wrong number.

One more thing belongs on the file before the truck moves, and it is about who carries the loss when something goes wrong. If the seal is intact and the count is short, the question is where the shortage happened, and the only evidence is the record made at each handover. If the seal is broken, the question is when it broke, and the only evidence is the photograph and the timestamp you took at each end. Neither of those is a legal position, and I am not going to tell you who is liable. What I will tell you is that the party with the better record wins these conversations, and on a transit file the record is four photographs, two timestamps and a number written down by hand. That is about fifteen minutes of work per shipment, and it is the cheapest thing on this entire page.

What I would do this week is short and it is not a strategy. Run one shipment through it, all the way, with your own timestamps at every handover, and compare your numbers against mine rather than believing them. Reconcile master and house before landing, every time, without exception. Write the seal number on your own file and not just on the operator's. Confirm free time and waiting charges in writing for the site you will actually use. Keep thirty per cent on the old path. Put a date in the diary to reprice the whole thing against the frequency that actually operated.

Twenty-eight point three tonnes is a small number. What it tests is whether a second door works, and a second door is worth more than the tonnage suggests.

↑ Back to the key points

— By Derek Xu