Energy price rises exacerbated by algorithmic energy trading

Europe’s gas market volatility, which has doubled during the Iran war, is being stoked by headline-reading energy trading algorithms. That is according to a new investigation from real-time energy newswire Montel News.

Computer programmes which are able scan headlines and social media posts, before using that information to trade energy, have been amplifying intraday swings on the Dutch TTF front‑month gas contract (Europe’s benchmark) since the start of the conflict in Iran.

According to Montel’s Chief Analyst, Tobias Federico: “Volatility has more than doubled on the front-month TTF contract since the outbreak of war on 28 February, with a fifth of the world’s LNG shipments stopped almost entirely.” 

One trader, one hundred bots

An individual trader could “easily have 100 to 200 bots following markets and placing bets”, for example, Professor Antonio Mele, a macroeconomist now doing AI research at the London School of Economics, told Montel.

“It is like having 200 traders working for you for cents on the dollar. They process news and execute orders automatically and that makes prices move much, much faster.”

Since the war in Iran began, all eyes have been on the Brent crude oil price when tracking market sensitivity to social media posts but increasingly European gas markets long anchored in fundamentals were exhibiting the same reactive dynamics. This has, for example, been prompted by the likes of US president Donald Trump’s daily dispatches, leading to large price swings that profit some traders, but leave others lamenting an “untradeable” market.

Speaking to Montel, Alan Whitefield, Trading Manager at Alcarbon said: “The swings in oil that we’ve seen relating to algos reading headlines, the TTF is following almost the exact same sort of patterns.

“It feels like the gas market reaction is much more violent and sudden than it used to be in previous crises.”

He added: “It’s just a self-fulfilling prophecy, if your algo is mirroring what others are doing then you are jumping on a trend.” 

Fake news and false information

Programmes now “score the sentiment” to rate the price-moving potential of short posts and headlines and send orders in milliseconds.

Many algo systems use blunt keyword triggers to start automated trades, such as “war”, “murder”, “closing of straits”, “bombing”, “end of civilisation”, Mirko Ivanda, a Geneva‑based cross‑commodity trader told Montel.

Ivanda estimated the algorithmic reaction was responsible for “probably 50%” of recent intraday volatility linked to the Iran war, way more than the single‑digit role they played in earlier crises such as the start of the Ukraine war. The TTF front‑month contract was previously identified as being heavily influenced by algorithmic trading in a 2023 probe by the Dutch ACM energy regulator.

These were based on past texts and price reactions, making trades vulnerable to an avalanche of often unreliable social media posts, however, he added.

Another trader said: “There’s no barrier for fake news or false information to be spread. Even unreliable information can move markets. Even if 80% of an account is rubbish, it doesn’t matter. If the market is convinced, it moves the market.”

Case study

Among “numerous” episodes, market participants pointed to a brief post on 24 April about possible Iran‑US talks.

At 13:08 CET that day, First Squawk, a global financial news service producing short posts every other minute, posted: “Pakistan may announce today the resumption of negotiations between Iran and America – sources.”

Within minutes TTF front-month volumes surged and prices plunged as bots rushed to sell positions on the bearish sentiment expressed in the post, Montel data showed.

The mechanics are straightforward. When lots of systems react to the same short post and speculators hold big one‑sided bets, a small push can set off a chain reaction.

Compounding the issue, traders said algos commonly took signals from related markets such as Brent crude, so a shock in one market quickly spread to gas.

More short-term could increase algo effects

The International Energy Agency warned in a recent report that the share of spot gas and flexible LNG in total EU supply could rise to around two‑thirds by 2030, which “could increase exposure to short‑term price volatility.”

The agency also noted that in the past two years the share of long‑term contracts together with domestic production had fallen from about 80% to roughly 50%.

That shift could help explain extreme positioning on the TTF front month, where the contract was currently trading around EUR 49/MWh at time of writing, up 53% since the start of the war. At the height of the conflict in late March, prices hit EUR 70/MWh, double what they were before the conflict began and a three-year high.

“The price swings are driven by multiple, compounding layers. It’s a volatility cake that explodes itself: each layer triggers the next, bigger volatility explosion. It is a real mess,” said Seb Kennedy, CEO and founder of market analysts Energy Flux.

“At the very best these markets are not functioning as intended and at worst are just broken and not fit to deliver accurate price signals,” he added.