Up to 30 Percent Fewer Missed Bags from IATA Resolution 753 RFID Tagging at Changi

Why a bag scanned four times at Changi can still go missing

A bag accepted at Singapore Changi Terminal 3 for a Paris-bound flight is scanned at four points before it reaches the hold. IATA Resolution 753, in force since 1 June 2018, obliges member airlines to record a bag at acceptance, at loading, at transfer between carriers, and at delivery to the arriving airport. Changi Airport Group built reader infrastructure across its terminals to satisfy that requirement, using UHF RFID inlays printed onto the same tag as the barcode. Changi runs that estate across Terminals 1 through 4, and its transfer volume makes the airport one of the higher-throughput test cases for the standard.

The transfer point at Changi generates the most dispute between carriers. When a bag moves from one airline to another, the receiving carrier’s departure control system has to accept a Baggage Processed Message from the sending airport. Where the two carriers run different DCS platforms, that message is translated in the middle, and translation is where events get dropped. A read at the transfer chute that never becomes a BPM leaves the bag invisible to the airline about to fly it.

Interline traffic through Changi is where this shows up first. A bag arriving on a Singapore Airlines service and departing on a partner carrier passes two separate systems, and the second one only sees the bag if the message arrives in a format it recognises. IATA publishes the message standards, but adoption of the newer event formats has been uneven across carriers, and the airport’s readers cannot compensate for a gap that sits on the airline side of the handoff.

What the 20 to 30 percent figure actually covers

Industry surveys put the global mishandled bag rate at roughly 6 to 8 per 1,000 passengers in recent years, and SITA has tracked the figure annually in its Baggage IT Insights series. Airport operators that deployed RFID at the four Resolution 753 points report reductions in the 20 to 30 percent band against their previous barcode-only baseline.

The terminology matters when reading the SITA figures. Mishandled covers delayed, damaged, pilfered and permanently lost bags, and the industry rate folds all four into one denominator. A missed bag is narrower: the bag did not arrive on the passenger’s flight. RFID tracking moves the missed-bag number more than it moves the damaged-bag number, because a tag read reports position and says nothing about condition. Airports with high transfer shares, Changi among them, carry a larger proportion of their mishandling cases in the transfer category than airports dominated by origin-and-destination traffic. Resolution 753 puts the obligation on the airline, while the reader estate at a hub airport is usually financed by the airport company.

Tag economics against the Montreal Convention cap

A few US cents per tag against a liability ceiling of 1,288 Special Drawing Rights per passenger is the arithmetic behind RFID adoption. RFID inlays cost several times what a printed barcode tag costs, though both stay in the range of a few US cents per unit at airline volumes. Industry surveys put the direct cost of a mishandled bag in the region of USD 75 to 150, which is the number that matters to a carrier deciding where to spend on tracking.

Rail, ferry and night train legs carry no bag record

Eurostar does not check baggage on any route. Passengers carry everything aboard, subject to a limit of two pieces of luggage plus one smaller item, and no tag, scan or bag message is generated at either end of the journey. Eurostar’s luggage policy is a carry-on policy with a size envelope, which means the operator has no bag record to lose in the first place.

Japan runs the opposite model on its high-speed network and still produces no shareable record. The Tokaido and Sanyo Shinkansen lines introduced a reservation requirement in May 2020 for luggage whose three sides exceed 160 cm, with a 1,000 yen surcharge on top of the fare. The Japan Rail Pass, which covers those lines, moved from 29,650 yen to 50,000 yen for a seven-day ordinary pass in October 2023. That is a 69 percent rise in one step, and it reset the arithmetic for anyone building a rail itinerary around the pass. A reserved oversized-bag seat is a booking, not a tracking event.

ÖBB Nightjet handles luggage the way a hotel handles it. The operator runs sleeper and couchette services linking Vienna, Munich, Venice, Paris and Brussels, and the passenger moves their own bags between the platform and the compartment. Nothing in a Nightjet booking generates a scan at boarding or at arrival.

Ferry routes work the same way, and DFDS, Stena Line and Brittany Ferries move foot passengers and vehicles with luggage staying with the passenger or in the car for the duration of the crossing. A booking reference identifies a vehicle or a foot passenger, and the bag travels outside both records.

Airport transfer costs in Tokyo show how far the intermodal chain extends without a bag record attached. The Keisei Skyliner from Narita to Ueno runs about 2,570 yen, the Narita Express about 3,070 yen, the Tokyo Monorail from Haneda to Hamamatsucho 500 yen, and the Airport Limousine bus between roughly 1,300 and 3,200 yen depending on which airport it serves. Each is a passenger movement with a ticket attached, and none of them carries a bag event.

A traveller who flies into Narita, rides the Skyliner to Ueno, boards a Nightjet in Vienna three days later and finishes on a Brittany Ferries crossing holds five transport records and one bag. Only the Changi record tracks a bag.

Where RFID reads fail

Reader gates at transfer chutes lose reads when a tag sits against a metal case frame or under a stack of hard-shell luggage. UHF inlays in the 860 to 960 MHz band, the range IATA RP 1740a tags use, are sensitive to the material immediately behind them, and a bag tag pressed flat against aluminium reflects rather than absorbs. Operators respond by adding reader antennas and by raising transmit power, which improves the read rate at the cost of more interference between adjacent gates.

A read event at a Changi gate also confirms less than it appears to. A tag passing a reader at the loading point tells the system that the bag was near the belt loader. It does not confirm that the bag went up the belt and into the hold. The gap between a scan and a physical load is the one that produces the bag recorded as loaded that then does not travel.

Changi’s planned Terminal 5 adds reader capacity without changing the underlying constraint. More gates raise the probability that a tag is read at each of the four points, and the message layer between the airport and the airline remains the place where a successful read can still fail to produce a usable record.

What the passenger can see, and what they cannot

Airline apps surface the last recorded scan point for a bag, which is the same event data the carrier receives from the airport. A bag shown as last scanned at Changi T3 acceptance, with no loading event after it, is a bag the airline cannot confirm went onto the aircraft. Passengers who report it at the arriving airport’s baggage services desk file a Property Irregularity Report, and that report opens a case in SITA WorldTracer, the tracing system most carriers use to match found bags to claims.

Under the Montreal Convention, a carrier can treat a delayed bag as lost once 21 days have passed from the date it should have arrived, and the liability limit of 1,288 Special Drawing Rights becomes the relevant number at that point. Until then the case sits in the tracing queue with the same event data the passenger can already see. Changi’s passenger-facing app shows the same four-point timeline the airline sees, so a traveller and a baggage services agent are often looking at identical records while disagreeing about where the bag is. Whether the next increment comes from cheaper inlays or from a shared message layer between carriers is, for now, an open question.