Credit risks and costs – and how the tide may be turning

Travelling offers time to read — and Buttonwood’s Into Thin Air column in The Economist, reflecting on receivables risks and fraud -  reminded me of questions I have asked for decades.

It all started when I was a young banker listening to a Helsinki taxi owner complain that my bank offered him a more expensive car loan than the car dealer.
As it happened, the dealer was also my client. I knew he had to borrow from the bank whenever customers didn’t pay cash — and he paid a rather high risk margin for that. Naturally, he then added his own margin and credit-assessment costs before offering finance to the buyer.

So why did it cost more for the buyer to borrow from the bank directly and pay cash? Rationally, it should have been cheaper — one risk assessment instead of two, and no cumulative margins.

This small example exposes a bigger structural flaw: long value chains where every supplier extends credit to the next. Each step adds its own risk costs. The end result is a hidden, systemic credit premium that raises prices throughout the economy.

If only the end-buyer took the loan — promising cash payments down the line — liquidity would flow instantly through the supply chain. Everyone would benefit.

But in the EU, political overreaction to some badly managed banks’ past mismanagement has made regulated credit assessments disproportionately expensive. That has pushed the SME sector more toward supplier financing — the least efficient and most risk-laden form of credit.

Now, however, the tide may finally be turning.
The emerging EU Trust Infrastructure — with verifiable credentials flowing through business wallets (EUBWs) — can radically streamline credit risk evaluation. Verified payment history, verified corporate identity, verified ownership and representation — all instantly checkable, all machine-readable.

This opens the way to fewer credit processes, fewer intermediaries, and lower accumulated risk margins. AI-agents, equipped with these verifiable data streams, will soon be able to price credit with unprecedented accuracy and transparency.

Buttonwood might want to revisit the theme from this angle: how verifiable credentials and AI-assisted credit analytics could deflate the massive, opaque layer of supplier finance — which even ChatGPT estimates to account for 5–10% of total credit volumes.