Decoupling Delusion

The future looks bright because economic growth requires ever smaller amounts of energy.

With lower levels of energy consumption and a growing economy, reductions in carbon dioxide and other environmental contaminants will occur naturally. Humans will not have to significantly modify the products we consume nor will we have to make major changes in our behavior.

That is the miracle of decoupling.

That is nonsense.

Decoupling is based on a belief that economic growth has required less and less energy consumption since the early 1970s. This is called the energy intensity of GDP.

The energy intensity of GDP is the amount of energy required to create a unit of economic growth measured by gross domestic product (GDP). This relationship is false and results from the fatal error of adjusting historical GDP for inflation without making a similar adjustment for the price of energy.

The Miracle of Decoupling

McKinsey captured the miracle of decoupling in its 2019 report Energy and GDP Growth.

The amounts of energy that economies need have increased virtually in lockstep with the amounts of wealth that economies create…Nonetheless, our analysis suggests that while a more populous world will create more wealth than ever, energy demand rates will plateau and demand rates for fossil fuels will begin to decline worldwide.

How can that be?

…A steep decline in energy intensity of GDP, primarily the consequence of a continuing shift from industrial to service economies.

The best of all possible worlds.

The correlation between oil consumption and economic activity (GDP) is well known (Figure 1, courtesy of Simon Michaux). Countries that use more oil have greater productivity and, therefore, higher GDP.

The miracle of decoupling arises when energy consumption and GDP are shown as a time series. Figure 2 shows the energy intensity of GDP—the same data shown in Figure 2—as a ratio for each year since 1970 (blue line) along with total energy consumption (orange line).

Morgan Stanley used a similar plot of the oil intensity of GDP to project that the world will use less than half the oil in 2040 to increase GDP $1000 as it did in 2000. Eureka!

The relationship is compelling but it violates most of the laws of physics.

Work is required to create the material production measured by GDP and work cannot occur without energy. How can the economy continue to grow by doing less work and using less energy? McKinsey says that the principal reasons are a shift from industrial to service economies and a marked increase in energy efficiency. 

These sound like reasonable explanations but clearly, the entire world can’t become a service economy. My colleague Thomas Murphy recently wrote,

Continued economic growth in the face of steady-state physical resources would require all growth to be effectively in the non-physical sector…Physical resources…must shrink to an ever-smaller fraction of the economy, translating to a small and diminishing fraction of an individual’s annual income having to go toward physical goods. All the food, energy and material purchases would become essentially free. This result makes little sense in the context of supply and demand.

Similarly, energy efficiency advances are real but infrequent and early gains are the most significant and easiest to make.

Efficiency improvements are not unbounded…Once a physical resource is saturated, we might expect some continuation of efficiency gains that can provide a modicum of additional economic growth. But it will probably be confined in both time and magnitude—the rate of improvement starting at less than 1% per year and declining from there.

Thomas Murphy

Fatal Error of Inflation Adjustment

Energy consumption is affected by the cost of energy. Analysts who make energy intensity curves adjust GDP for inflation without making an adjustment for energy consumption.

Energy was purchased using the money of the day, not an inflation-adjusted amount that was much higher. This progressively distorts the resulting calculation of  energy consumption per dollar of GDP for each year before the benchmark year for inflation adjustment.

Table 1 compares world GDP adjusted to 2020 dollars and using nominal GDP in the dollars of the day. It also shows global energy consumption and the resulting energy intensity of GDP values.

By inflating 1970 GDP to $27 trillion instead of the nominal GDP of $4 trillion in that year, the resulting intensity value of 2.42 is almost seven times lower than the 16.12 using nominal GDP.

Figure 3 graphically shows how much distortion is introduced into the resulting energy intensity curves.

The implications are profound. Rather than a linear decrease over time, the rate of change in energy intensity is now approaching zero.

Efficiency and shifts toward service economies play a part in decreasing energy intensity but these effects are over-stated. The sharp decreases in intensity before 2000 resulted more from productivity gains than from efficiency gains as more of the world shifted energy consumption to oil from coal and biomass energy sources.

The likelihood of future decreases in energy intensity are low. That means that it is improbable that future GDP will increase without substantial increases in energy consumption. That is bad news for climate change and for earth’s ecosystem.

When things look too good to be true, they usually are.