Inconsistent definitions of GDP: Implications for estimates of decoupling

Efforts to assess the possibilities for decoupling economic growth from resource use and negative environmental impacts have examined their historical relationship, with varying and inconclusive results. This paper shows that ambiguities in the historical measurement arising from definitional changes to GDP are sufficiently large to affect the results. I review the history of structural revisions to GDP using the example of the United States, and on international comparisons of purchasing power parity, and compare decoupling results using GDP vintages re -ported between 1994 and 2021 for most countries. Between vintages, 10-15% of countries switch between relative decoupling and recoupling from energy or materials on decadal intervals, and up to as many countries as decouple absolutely in an older vintage stop or newly start absolutely decoupling in the newer vintage. GDP vintages also affect environmental Kuznets curve results on absolute decoupling in Grossman and Krueger's seminal paper and accelerate the International Energy Agency's annual global decline in energy intensity by up to-0.2 percentage points. Inconsistencies in economic measurement introduce ambiguity into historical decoupling evidence and model projections into the future. To advance debate, rigorous reporting and sharing of data vintage for subsequent comparison and replication are urgently needed.