“Decoupling” refers to two curves— greenhouse gas emissions and economic growth (that is, changes in real gross domestic product, or GDP)—and the question is whether they can diverge, or even move in opposite directions: Can we see the real GDP curve rise over time, while the emissions curve rises more slowly (“relative decoupling”) or even declines (“absolute decoupling”)? We consider real GDP here because it allows for comparisons across different years by accounting for inflation (unlike nominal GDP). It is always real GDP that is discussed when analyzing time series.
Most often in the debate on decoupling, the focus is on the issue of greenhouse gas emissions (GHG) or energy—and thus solely on the climate problem. However, this is just one of the nine planetary boundaries currently identified by the scientific community. Theoretically, there are as many debates on “decoupling” as there are environmental indicators whose trends over time are being monitored: energy consumption, raw material extraction, overall environmental footprint, etc.
It should be emphasized that each of these environmental indicators is subject to questions regarding the relevance of the measurement, its reliability, and its methodology. Scientists are seeking to determine whether we are accurately measuring what we intend to measure and whether the phenomena are properly captured by the statistics. For example, when we refer to a country’s emissions, do we mean those associated with what is produced within its borders, or those associated with what its residents consume—which includes emissions from imported goods?
GDP, on the other hand, is always taken for granted. The use of this indicator is rarely questioned. Yet to what extent do GDP data provide an “accurate” representation of our economies? The fact is that when we take into account the statistical uncertainty associated with “properly” measuring economic activity, the margin of error in identifying a “decoupling” increases. And with it comes greater uncertainty or risk-taking associated with “green growth” strategies, as opposed to paradigms of frugality, post-growth, or degrowth. The article recently published by Gregor Semieniuk addresses this very question: Are we measuring economic activity correctly, or, to put it another way, is GDP a reliable indicator for this purpose?
Several critical steps can be identified in the calculation of GDP that could lead to uncertainty regarding its value in the context of the decoupling debate. First is the question of scope. Over time, accounting conventions have expanded the scope of economic activities included in the calculation, often for technical reasons but also for sociopolitical ones. For example, financial activities were not added to the national accounts calculation guidelines until 1968. Prior to that, they did not “count” toward the indicator, as they were considered unproductive. Similarly, in 1977, services provided by public administrations were included in the scope of GDP, reflecting significant changes, particularly in the view that public activities generate wealth.
Second, some outputs do not have a market price, and their value is, by convention, equated with their cost of production. This is precisely the case for these nonmarket services provided by public administrations. An output provided by a public service is thus generally worth less than the same output provided by a private firm, since the cost of production is lower than the market price, which, in turn, includes profit.
Third, and this is an absolutely crucial step, “real” GDP—also known as “volume” GDP—is calculated by adjusting it for price changes to allow for analysis over time. In France, INSEE remains relatively tight-lipped about its methods for calculating inflation. For example, the goods and services that make up the basket on which the price index is based are always kept secret. This is largely the result of significant political stakes and pressures surrounding its value: readers can easily imagine the incentive a government might have to report low inflation when social benefits, pensions, or minimum wages are indexed to it.
The very philosophy behind calculating inflation has changed drastically over time, shifting from an indicator representing a “typical” working-class family in the Paris region to one that aims to represent the cost of living for a theoretical “average” consumer. Each of these approaches has its own legitimacy, but it is important to recognize that inflation, as currently defined, measures a kind of cost of living that, strictly speaking, is not experienced by anyone.
In addition to the composition of the basket of goods and services, the calculation of inflation is also subject to conventions regarding how to account for changes in the composition of the basket, particularly changes in quality. Debates continue regarding the “correct” way to measure it, and some estimates vary by as much as a factor of two internationally, often on the lower end.
Let’s really think about what this means: if inflation is X percentage points higher or lower, then the deflated GDP—the figure we use every day and for all historical comparisons—contains an error of the same magnitude each year, and thus cumulatively! Not to mention the problem of relative prices, which change over time and significantly alter the values of historical time series depending on the reference point.
In this discussion of the degree of confidence that can be placed in GDP in the debate on decoupling, Gregor Semieniuk’s work is, to the best of our knowledge, the first to examine the impact of the various definitions of GDP—developed over time—on decoupling results. The author thus traces the various GDP series proposed over time and the “structural revisions” adopted, which concern the calculation method, its scope, or the base year for inflation.
The trend is clear: the more recent the definitions, the higher the current GDP, and the stronger the past growth (the graph presented earlier in the article, taken from Semieniuk’s study, illustrates this for the 1978 and 2018 definitions). This, however, raises the daunting question of which is the “right” definition for understanding our economic history: the 1950 definition or the 2020 definition? The 1950 definition was undoubtedly considered more relevant at the time for describing the economy. Each version is heuristic at the time it was developed.
However, the results of “decoupling” or “recoupling” vary greatly depending on the GDP series used. With the recent definitions, historical growth is stronger than with a more “industrial” definition of GDP, and it is therefore easier for the associated curve to diverge from that of environmental impacts. This second graph, also taken from Semieniuk’s study, illustrates this point using energy as an example: in one case, the statistics paint a picture of an economy that requires 50% less energy to produce a unit of wealth. In the other, progress has been only 30%. The only difference is the definition of the economic indicator—GDP.
This allows us to compare the various “decoupling” results based on how GDP has been defined over time. By simply changing the measure of economic activity—which we usually take for granted—Semieniuk transforms decouplings into recouplings for certain countries, and vice versa (!). The older the definition, the more countries shift into a decoupling situation; however, on average, 10 to 30 countries shift from one decoupling state to its opposite when the GDP definition is changed—a phenomenon that can be considered a statistical artifact.
An additional layer of confusion arises when comparing different countries. Ideally, GDP time series should be comparable across countries and should not be influenced by differences in national currencies. The idea is that the same level of GDP per capita represents the same standard of living—that is, access to the same “standard” set of goods and services .” To achieve this, we use “purchasing power parity” (PPP) data, the methodology for which is even more complex than that used to measure inflation: how can we rigorously compare “purchasing power” across all countries in the world, given that consumer cultures are not easily comparable?
Once again, we find various measures whose definitions have evolved over time. And when analyzing data for the same country, the growth rates and the value of GDP in PPP—in their various forms—sometimes yield results that differ significantly. In principle, the PPP approach is more accurate for international comparisons than a simple conversion of national currencies into dollars, since exchange rates fluctuate from year to year and are sometimes driven by purely speculative factors. However, it requires a certain degree of methodological finesse.
This leads us to another consequence, particularly for countries where statistics are unreliable: if the growth rates of countries in the Global South are over- or underestimated, this implies major changes in the credibility of the emissions reductions required of them in climate change mitigation scenarios. An overestimated GDP paints a very optimistic picture of past trends in environmental performance relative to economic performance, and these trends will be extrapolated in the transition models used, in particular, by the IPCC.
Finally, let us recall that Gregor Semieniuk’s study addresses only one of the various aspects of the definition of GDP (that of relative prices). Raising the question of the credibility of GDP calculations—and, above all, that of its necessary yet so often neglected interpretation—opens the door to further scrutiny of how we conceptualize the progress our economies have made, and the path that still lies ahead toward mitigating climate change. Its significance, so often presented as self-evident, is in reality a delicate issue. More generally still, the question arises as to what GDP growth can truly tell us about the health and trajectory of economic activity. To what extent are we simply telling ourselves stories?
