The rise and rise of confirmations in commodity trading
Tuesday, April 5, 2022
(Physical trading books have trade entry error rates systematically greater than 15%. You just don't know it yet)
A commodity transaction is a promise to deliver a specific grade of the commodity at a given location, at an agreed date in the future. Physical trading is therefore an exchange of promises, where typically buys and sells can be offset (and sometimes “booked out”) if the parameters of the trades (grade, location, delivery period) are sufficiently generic.
Once a commodity transaction is executed, there are usually three “pain points” to make good on its underlying promise: a contract must be written as quickly as possible after the trade, to remove any potential misunderstanding; the title of ownership of the commodity “cargo” needs to pass smoothly from the seller to the buyer at the time of delivery; and the agreed price needs to be calculated (when it is not a simple fixed price). Paying for the “cargo” is usually fast and efficient, although the act of financing it has remained quite painful. The quicker and the cheaper these processes are, the lower the risks in trading books, and the higher the churn of the market.
Today we will focus on the contracting phase and its two competing post-trade infrastructure models: “paper contracts” and “electronic confirmations”. In historical markets, traders sign General Terms & Conditions (GT&Cs) and then document each transaction with a “pdf contract”, usually dealt with by a contract officer, or a “physical operator” also in charge of the optimisation of the portfolio of contracts and the transfer of title for cargoes. This traditional method of coming to a legally binding agreement is the one that prevails in physical oil, LNG, metals, concentrates, and agricultural commodities. Each “pdf contract” needs to be read by a human, which is very inefficient given that these contracts are often supposed to be generic and easily tradable. In physical oil trading, it is common for the buyer and the seller to “agree to disagree” on clauses triggered infrequently such as sanctions and to decide not to execute the contracts, leaving them in a limbo state . Trade finance bankers point out that 90% of the trades they are asked to finance are not “papered”. More importantly, at VAKT, our recent experience in ARA Barges based on algorithmic reconciliation of a large volume of trades is that the error rates are always higher than 15%. You’ve read that right: if you manage an oil trading book, at least 15% of the trades in your book are wrong, either because your traders mis-booked them, your counterparty mis-booked them, or you fundamentally disagree on some aspects of the trades with your counterparty. This is a sobering perspective.
By contrast, more recent markets, such as those that emerged in European and US Gas & Power in the 1990s, are organised around Master Trading Agreements (MTA) and confirmations. With an MTA, the traders agree most of the generic terms of the transactions at the outset of their trading relationship, and they only confirm the commercial details at the time of execution. These commercial details are codes (representing reference data), units and numbers that are easily bundled into a “digital payload” and sent to a “electronic confirmation facility”, which is an on-platform pairing and comparison algorithm. The MTA between the parties specifies that a match of the commercial details on the confirmation facility is as good as signing a “paper contract” and – voilà! – “pdf contracts” disappear.
Confirmations are addictive. They lead to dramatically lower error rates in trading books – typically less than 0.5% – and, through instantaneous feedback loops for careless traders, to better discipline at trade entry. Middle officers only need to attend to the rare exceptions which have not matched automatically. Confirmations also help normalise reference data across the marketplace and allow for all stakeholders to describe the commercial details of transactions with a shared language. With some further contracting between the parties, they typically lead to netting and “book outs”, reducing market and credit risks in trading books. And at a macro level, they help increase the churn of the markets by making them more efficient and cheaper to operate.
There are good reasons why confirmations have taken their time in physical oil. Whilst it takes a handful of fields to confirm a UK natural gas transaction for instance, it typically takes 50 to confirm a refined product transaction, with a total of 135 atomic data fields when all the details of the delivery terms and of the price formulas are spelled out in excruciating detail.
Finally, a critical point that is often missed is this: if you can confirm a trade (with software), you can also execute it “electronically”. Historically, marketplaces have tended to pre-date confirmation facilities: it was the sheer volume of electronic transactions and the clogging of back-offices that justified the need for automated contracting processes. I remember fondly the “paper confirmations” of Gas & Power and oil derivative markets of the 2000s; they did not last long. It can be the other way around too: credit derivatives, whose complexity easily rivals that of physical commodities, saw the digitisation of confirmations first, which opened the way to the creation of a marketplace for credit default swaps by ICE. It will be interesting to see which path our physical commodity markets will follow to become fully “electronic”. All markets do, eventually.
At VAKT, we are excited to be working on the first confirmation facility for physical oil trading. Indeed, our clients are already amending their bilateral trading agreements to enable digital confirmations on the VAKT platform, and we expect the first confirmations soon. Once established, these digital confirmations will vastly reduce the operational risk profile of our customers, increase their profitability, and lead to happier teams who are freed from the burden of the manual contracting process.
 There is even a weird notion that the last party to have replied to an email would prevail in a dispute. As would be expected for generic market instruments, litigation is carefully avoided between large market counterparties: claims on contracts get bundled together and settled regularly between the traders.
 The same is true for non-commercial details through market-standard MTAs.
 Nowadays, such an execution is more likely to be “optical” than “electronic” though.