It is widely assumed that over the coming decades, increased energy efficiency will help the world meet its energy needs and reduce carbon emissions. That may be true, but recent research suggests that energy intensity — a widely used way of measuring efficiency — isn’t the right metric.

Energy intensity is a simple ratio: energy use per dollar of GDP. According to the U.S. Energy Information Administration, the energy intensity of the U.S. economy declined by 1.6 percent per year from 1985 to 2004, suggesting that we’re doing more with less energy.

Energy companies have embraced this idea: “Energy per unit of income as measured by GDP continues to fall, and at an accelerating rate,” asserts BP. “This is true in our outlook to 2030 not only for the global average, but for almost all of the key countries and regions. The combination of energy efficiency gains and a long-term structural shift towards less energy intensive activities as economies develop underpins this trend.” Similarly, ExxonMobil’s “Outlook for Energy: A View to 2040” projects that economic output will increase by 80 percent while OECD energy demand remains flat, because “[g]lobal energy demand does not rise as dramatically as economic growth as a result of declining energy intensity.”

That sounds pretty good, and to a certain extent it’s true. Numerous changes in the U.S. economy have gradually made it more energy efficient since the 1970s, and it is widely observed that the world’s most developed economies are more energy efficient than developing economies. However, it’s far from certain that developing economies will achieve enough wealth to attain First World levels of energy efficiency, as agencies like the Intergovernmental Panel on Climate Change assert, or that the historical trend of improving energy efficiency will persist into the future, as many other observers, like Credit Suisse, assume. If these assumptions don’t pan out, we could encounter a climate and energy supply catastrophe much sooner than imagined in official scenarios.

Energy intensity of the U.S. economy.
Credit SuisseEnergy intensity of the U.S. economy.

First, the very structure of the energy intensity ratio is problematic. As the U.S. Energy Information Administration notes, it “overstates the extent to which energy efficiency improvements have occurred in the economy.” For one thing, when the economy is growing faster than energy consumption — when the GDP denominator is growing faster than the energy numerator — energy intensity falls. But in a recession, GDP tends to fall faster than energy consumption, and so the energy intensity ratio rises. This would be true even if there were no change in energy efficiency. Indeed, the Paris-based International Energy Agency notes in its “World Energy Outlook 2012” that in 2009 and 2010, global energy intensity actually increased due to economic contraction, lower energy prices (which boosted energy consumption), and colder or hotter weather in certain regions. “The production of a huge number of goods, the mixing of the transport of freight and people, and the variety of housing and climates makes an aggregate energy intensity number based on Gross Domestic Product (GDP), a number that disguises rather than illuminates,” the Energy Information Administration says.

Energy intensity is also affected by population size, travel patterns, natural resource endowments, changing activity within the economy like shifting from steel production to electronics, technological improvements like using robots, and off-shoring of energy-intensive industries like manufacturing to other parts of the globe. For the past several decades, the United States and Europe have obtained increasing amounts of their manufactured goods from Asia as they transitioned to service-oriented economies, hiding some of the actual energy they consume. This has led to accusations that industrialized countries have “cheated” on their decarbonization claims by substituting imports for domestic production in energy intensive industries. There may be something to those claims.