The data shows that there are excess deaths beyond what can be directly attributed to Covid-19, that Covid-19 deaths are themselves not strongly correlated with

The data shows that there are excess deaths beyond what can be directly attributed to Covid-19, that Covid-19 deaths are themselves not strongly correlated with energy prices, and that excess deaths in 2022-2023 were moderately correlated with energy prices. So we can be somewhat confident that Covid-19 is not the primary causal pathway to excess deaths in 2022-2023. The worst problem is that the author concluded that there is a causal path between the high prices and excess deaths, which they absolutely did not do enough work to justify. For example, I don't see any longitudinal analysis here, nor do I see the effects of weather (temperature, humidity, cold snaps, high winds) accounted for. If your main causal hypothesis is that physically being cold caused the excess deaths, then something like 5th percentile temperature, events like "cold snap below -5 C for 3 or more days", etc. should be strongly correlated with excess deaths, after controlling for energy prices and baseline seasonality over several years. Consider that temperature might actually be the underlying causal factor for both high prices and excess deaths! That is: cold winter causes more deaths, and cold winter also causes higher prices. Maybe also having a warmer climate is correlated with being less dependent on Russian gas for heat due to geopolitical factors. That would be an interesting article in its own right, but it's not quite as exciting of a headline. These points should be incredibly obvious to anyone writing for The Economist, and the fact that this wasn't even mentioned makes me think this article is more than just "incompetence" but outright bad faith. > As wholesale energy prices fall and temperatures rise, the immediate threat may be over, but it is clear Mr Putin’s energy weapon was deadly. Ugh, does it get any cornier? I feel like readers ought to start demanding data and source code along with allegedly "data-driven" articles like this. That way at least I can do my own analysis using theirs as a starting point. Their list of sources at the bottom is not exactly something I can work with: > Chart sources: The Economist’s excess-deaths tracker; Copernicus; HEPI In any case, junk articles like this are exactly the reason I haven't subscribed to The Economist in years. Blogs written by actual economists seem to be the only place to find any sober analysis nowadays.