For eighty years, the answer to “will I actually get paid?” belonged to a bank. Under UCP 600, a documentary credit is the issuing bank’s own undertaking: the exporter ships, presents conforming documents, and the bank pays because the bank promised to. The buyer’s creditworthiness matters, but it is the bank’s problem, priced once when the credit line is set.
The International Chamber of Commerce moved that promise. In the Uniform Rules for Digital Trade Transactions (URDTT), published in October 2021, the central instrument is not a bank undertaking at all. Under URDTT, a Payment Obligation is an irrevocable obligation incurred by the buyer — a definite undertaking to pay at sight, or on a fixed or determinable future date, to the beneficiary.
Read that again, because it is not a technical detail. The rules that describe digital trade have quietly relocated the promise from the balance sheet of a regulated bank to the balance sheet of your customer.
UCP 600 priced a bank. URDTT prices a buyer.
This is the shift that matters, and it is easy to miss inside the broader digitisation story. Most of the conversation about URDTT is about what it removes: paper, presentation, examination, the discrepancy cycle. The rules are deliberately technology-neutral and messaging-standard-neutral, and they are designed to sit alongside the UNCITRAL model laws rather than fight them. All good.
But removing the document does not remove the risk. It relocates it. In a documentary credit, an exporter in Osaka does not need a view on a buyer in Lagos, because a bank has taken that view already and sold it to them as an undertaking. Under a buyer’s Payment Obligation, that intermediation is gone. The exporter — or whoever finances the exporter — is now looking directly at the buyer and asking the oldest question in trade: is this company good for it?
URDTT also opens the door deliberately wide. It was written for corporates and for the broader spectrum of financial service providers, not for banks alone. That is the right call commercially. It also means the party carrying the buyer risk is increasingly not an institution with a credit department, a country limit and a hundred years of loss data.
URBPO already failed once.
We have run this experiment. The Bank Payment Obligation, governed by the URBPO, was the previous attempt at a data-driven substitute for the documentary credit. It did not fail because the technology was wrong or because the data matching did not work. It failed, in the words of the people who built its successor, because it was written for banks — and corporates wanted the payment obligation in their own favour, not in the bank’s.
URDTT fixes precisely that. The obligation now runs to the beneficiary, which is what corporates asked for. It is a genuinely better rulebook, and the ICC deserves credit for listening to the market rather than re-issuing the same product.
But fixing who holds the promise does not answer who underwrites it. A rulebook can define an irrevocable buyer obligation with perfect clarity and still leave the exporter holding an obligation from a company nobody will price. The instrument is only as financeable as the credit assessment sitting behind it.
The file that would price the buyer sits with the seller.
Ask an underwriter how a buyer limit actually gets set and the list is consistent: audited financials, payment experience, credit agency reports, public records, company registries, market intelligence, adverse news, sector conditions. It is a careful process and a genuinely skilled one. It is also almost entirely third-party and backward-looking.
For a mid-sized importer in Germany or Japan, that produces a decent picture. For an SME buyer in a thin-file market — much of Southeast Asia, most of Sub-Saharan Africa, large parts of South Asia — it produces almost nothing. No filed accounts worth reading, no agency file, no rated peers. The underwriter is not being unreasonable when the limit comes back small or empty. They are being asked to price a company that is, on the available evidence, invisible.
Meanwhile, one party knows exactly how that buyer behaves. The exporter has been trading with them for four years. They know the buyer has taken delivery on forty consecutive shipments, paid on day 58 against 60-day terms, disputed twice and settled both, and increased order size every quarter through a currency shock. That is the single most predictive dataset in the transaction, and it is the one dataset that never reaches the underwriter in a form anyone will accept — because it lives in the seller’s own ERP, in their inbox, in a spreadsheet, carrying the obvious objection that the seller is not a neutral party.
So the architecture moves and the limit does not. URDTT hands the exporter a buyer’s promise. The market still cannot tell them what that promise is worth.
This is the layer we are building
Standage’s position here is narrow and specific. We are not a credit insurer and we do not set limits. What we operate is the layer underneath: the shipment-level record of what actually happened between a specific buyer and a specific seller — orders, documents, delivery, and the timing of payment — captured as it occurs rather than reconstructed afterwards by the party with an interest in the answer.
That record is not a credit decision. It is the evidence a credit decision has never had. The work we are building toward is making it portable and verifiable enough that an underwriter, a bank or a non-bank financier can rely on it without simply taking the seller’s word for it — the same shift that made electronic bills of lading bankable, applied to payment behaviour rather than to title.
We are early in that, and we would rather say so than overstate it. But the direction is settled by the rules themselves: if the promise now belongs to the buyer, then the buyer’s trade record has to become something a third party can price.
The litmus test
Put one question to your bank, your insurer or your financier this quarter:
“My buyer has taken delivery and paid me on time on forty consecutive shipments. Does that raise their limit?”
If the answer is yes, your trade record is already doing financial work and you should be pushing far more of it into the file. If the answer is no — and today it usually is — then your trade is being priced on somebody else’s data about your customer, while the best evidence in the transaction sits unused on your own systems. That gap is not a paperwork problem, and no amount of digitisation closes it. It is the next thing that has to be built.
Related reading
- Everyone is digitising the paperwork. Most of world trade doesn’t run on it.
- Thirty organisations touch one shipment.
- Two banks settled across a blockchain. Settlement is still where trade breaks.
約束は銀行から買い手へ移った。与信ファイルは動いていない。
80年にわたり、「本当に代金を受け取れるのか」という問いに答えてきたのは銀行でした。UCP600のもとでは、信用状は発行銀行自身の支払確約です。輸出者は船積みし、条件に一致した書類を呈示し、銀行が約束したがゆえに銀行が支払う。買い手の信用力は問題ではあるものの、それは銀行の問題であり、与信枠を設定する時点で一度だけ値付けされます。
国際商業会議所（ICC）は、その約束の置き場所を動かしました。2021年10月に公表された「デジタル貿易取引統一規則（URDTT）」において、中心となる仕組みはもはや銀行の確約ではありません。URDTTにおける支払義務（Payment Obligation）とは、買い手が負う取消不能の義務であり、一覧払い、または確定した将来の期日に受益者へ支払うという確定的な約束です。
UCP600は銀行を値付けした。URDTTは買い手を値付けする。
これは技術的な細部ではありません。書類、呈示、点検、ディスクレの往復——URDTTが取り除くものに議論が集中しがちですが、リスクは取り除かれたのではなく、移動したのです。信用状であれば、大阪の輸出者はラゴスの買い手を自ら評価する必要がありません。銀行がすでにその評価を行い、確約という形で売っているからです。買い手の支払義務ではその仲介が消え、輸出者（あるいは輸出者に資金を出す者）が買い手を直接見て、貿易における最も古い問いを立てることになります。この会社は支払えるのか、と。
URBPOは一度失敗している。
この実験はすでに行われました。URBPOに準拠する銀行支払確約（BPO）は、信用状に代わるデータ駆動型の仕組みとしての前回の試みです。失敗の理由は技術でもデータ照合でもありません。その後継を作った人々の言葉を借りれば、銀行のために書かれていたから——事業会社は支払義務を銀行のためではなく自らのために求めていたからです。URDTTはまさにその点を修正しました。しかし、約束を誰が持つかを正したことは、それを誰が引き受けるかの答えにはなりません。
買い手を値付けできるファイルは、売り手の手元にある。
与信枠がどう決まるかを引受人に尋ねると、答えは一貫しています。監査済み財務諸表、支払実績、信用調査会社のレポート、公的記録、商業登記、市場情報、ネガティブ情報、業界環境。周到で、熟練を要する作業です。同時に、そのほとんどが第三者由来であり、過去を向いています。
ドイツや日本の中堅輸入者であれば、それで相応の像が結べます。しかし東南アジアの多く、サブサハラ・アフリカの大半、南アジアの広い範囲といった「ファイルの薄い」市場の中小企業では、ほとんど何も出てきません。引受人が小さな枠しか出さない、あるいは枠を出さないのは不合理ではないのです。入手可能な証拠の上では、その会社は事実上見えていないのですから。
一方で、その買い手の振る舞いを正確に知っている当事者が一人います。輸出者です。4年取引し、40回連続で引き取りが行われ、60日サイトに対して58日目に入金され、2度のクレームがいずれも解決し、通貨ショックを挟んでも四半期ごとに発注量が増えてきたことを知っている。取引の中で最も予測力の高いこのデータは、引受人が受け入れられる形では決して届きません。売り手のERPの中、受信箱の中、表計算の中にあり、しかも「売り手は中立ではない」という当然の反論が付いて回るからです。
私たちが作っている層
この論点におけるSTANDAGEの立ち位置は限定的です。当社は信用保険会社ではなく、与信枠を設定する主体でもありません。私たちが運用しているのはその下の層——特定の買い手と特定の売り手のあいだで実際に起きたこと（受発注、書類、引き渡し、そして入金のタイミング）を、後から利害関係者が再構成するのではなく発生時点で捉えた、船積み単位の記録です。
これは与信判断ではありません。与信判断がこれまで一度も持てなかった証拠です。売り手の言い分をそのまま信じずとも引受人・銀行・ノンバンクが依拠できるだけの可搬性と検証可能性を備えさせること——電子船荷証券を銀行が扱えるものにしたのと同じ転換を、権原ではなく支払行動に対して行うこと。それが私たちの目指す方向です。まだ途上であり、そこは誇張せずに申し上げます。
リトマス試験
今四半期、取引銀行・保険会社・金融提供者に一つだけ問いを投げてみてください。「この買い手は40回連続で引き取り、期日通りに支払っています。それで与信枠は上がりますか」
答えが「はい」なら、あなたの取引記録はすでに金融的な仕事をしています。もっと多くをファイルに送り込むべきです。答えが「いいえ」なら——そして現状はたいてい「いいえ」です——あなたの取引は、自社の記録ではなく他人が持つ顧客データで値付けされています。最良の証拠は自社システムの中で眠ったままです。これは書類の問題ではなく、デジタル化を進めても閉じません。次に作られるべきものは、ここにあります。
