There are no rich, low-energy countries

Take every country that reports both figures, once per year from 1980 to 2023: you have 8,090 economy-year observations of real GDP per capita against primary energy consumption per capita. Primary means all energy use: electricity, transportation, industrial production, everything.

This is not the claim that growth cannot be separated from energy. It is a narrow empirical claim: in the joint distribution of income and energy, the corner marked high GDP, low energy, is empty. Over four decades and two hundred countries, no economy has yet been rich while consuming little energy.

Fitting the data

Both variables span three to four orders of magnitude, so we should use log–log scaling. Running standard OLS and plotting the mass, it collapses onto a tight diagonal band. The correlation of $\log E$ with $\log G$ is 0.90, and a typical country sits within a factor of about two of the central line.

The lower-right corner is empty. To make that precise, take at each income level the 5th percentile of energy use: the least energy that any country of that wealth has ever run on.10,000 kWh/yr is roughly a continuous 1.1 kW draw, about eleven incandescent bulbs per person. That is the floor, not the average. Less wealthy countries should consider this minimum level of energy consumption as a goal for their country if they are to increase their wealth. This lower envelope is essentially a lower bound on requisite energy consumption for your GDP, and it climbs monotonically: raise the income floor and the energy floor rises with it, with no exceptions across the entire range.

The most energy-frugal rich economy in the data is Macao—a casino-and-services city-state with essentially no industry, no agriculture, and no hinterland, which imports every joule-intensive thing it touches. The places that press hardest against the wall are exactly the ones typical economic theory predicts: financial and tourist enclaves that have outsourced their steel, cement, and freight to somewhere else. Even they only touch the floor. None has gone through it.

Change within countries over time

The pooled correlation combines two questions. The fact that richer countries are hungrier for energy than poorer ones is obvious. The temporal question is the one that matters for anyone projecting a country’s future: when a single country’s own income rises, does its own energy use follow?

Split each series into a country’s long-run mean and its deviations from that mean, and read the answers off separately:

• Between countries (long-run means): $r_B = 0.91$. Richer nations use more. No surprise.
• Within a country over time (deviations from its own mean): $r_W = 0.68$, slope $\beta_W = 0.80$. When a country grows, its energy use grows with it—sub-proportionally, because efficiency improves and output tilts toward services, but it grows. A 1% rise in real income has historically come with roughly 0.8% more energy.

Both slices of the data agree. The empty lower-right corner is not an artefact of comparing frozen snapshots of two hundred different societies; it holds along each society’s own path through time, even as countries have become steadily more efficient.

But is the converse true?

The lower-right is empty; the upper-left is notPetrostates like Turkmenistan and Trinidad & Tobago sit here; they own abundant cheap energy, and are middle-income; which is exactly why energy is necessary but not sufficient., which means the implication runs strictly one way. Abundant energy is necessary for prosperity but not sufficient.

For the time variable, it’s obvious that economy-years are not independent draws. A single economy contributes up to 44 highly correlated points, so the effective sample is far smaller than 8,090. But the empty region that stays empty through the fall of the Soviet Union, the rise of China, the internet explosion, and many wars, strongly suggests this is structural.

In the light of geopolitical events such as the Ukraine and Iran wars, and forward-looking events such as the AI revolution, this should alarm politicians in countries such as the United Kingdom, which has allowed its energy costs to skyrocket. Too often the recent spike in energy costs across Western economies is framed as a “Cost of Living Crisis”; with consumption typically closely-aligned with price, this has the potential to enormously underestimate the importance of the issue.

Causality can run both waysThe direction of causation is much debated and is named the “energy–growth nexus”. See Kraft & Kraft (1978) and the survey by Ozturk (2010)., but it doesn’t matter for the statement, which is all the title asserts: no country has yet been rich without commanding a great deal of energy. Anyone who claims the future will be different is entitled to their claim, but they are claiming something with zero precedent in the modern record, and the burden of naming the mechanism is on them. Until then, the conclusion the data will bear is this: cheap, secure, abundant energy is not a luxury that prosperity affords. It is a precondition for it.