Fuel-Shock Transmission into European Power Prices
GB, DE, and the Limits of Observational Inference after the February-March 2026 Oil and Gas Shock
Modo Energy has offered excellent analysis of the current gas/oil crisis on power markets: if gas remains marginal in peak and shoulder hours, then a sharp gas shock should pass into short-term power prices, though perhaps more weakly in Germany if coal and lignite still cap part of the move.
I wanted to investigate this logic further with open data so I used the Stratnergy analysis engine to investigate the February-March 2026 fuel shock as an observational transmission problem.
After the same geopolitical oil and gas shock, where did repricing appear most clearly in Great Britain and Germany, and what level of causal confidence can open public data actually support?
The exploration is fascinating and continues to develop. I summarise where I am with it here with a more detailed note in the terminal. Hopefully this prompts further ideas for how we can think about this situation and what we can learn.
The picture is asymmetric. GB provides the clearer same-day association pattern in the available short-term benchmark objects, but the British reading rests on a consistent pattern rather than on any single coefficient: same-day linkage, positive day-to-day change alignment, lag-zero dominance, and broader balancing-stress layers moving in the same direction.
Germany’s transmission path is more filtered. Daily-mean linkage weakens materially under adjustment and is more vulnerable to future-lag persistence. We therefore introduce a narrower screen than simple gas output: in how many evening intervals is gas still needed in the system and still close enough in cost to coal or lignite to be competitive inside the thermal pricing range?
In the broad pre-shock versus post-shock comparison, the German 18:00–20:00 premium over 12:00–16:00 rises while the gas-in-pricing-range share falls as gas moves above both hard-coal and lignite variable cost. The gas-in-pricing-range share falls further to 0.000 in the later escalation block. If gas was still setting price in many of those intervals, additional constraints that open public data do not observe—including unit constraints, CHP obligations, ramping needs, location, start-up limits, or balancing roles—would need to do more of the work.
Britain carries the stronger conditional-association case, whereas Germany carries the stronger timing-and-mechanism case. In Pearl’s terms, the contribution is not intervention-level identification, but disciplined separation between descriptive shock facts, conditional association, lead-lag diagnostics, and mechanism-consistent timing.
See stratnergy.online for more, where content continues to develop and build over time.