Beyond Aggregation Bias: The Importance of Consistency with Official Statistics in MRIO-Based Carbon Footprints

To understand how much pollution a country is responsible for, scientists often look beyond the smokestacks and factories within its borders. They ask a different question: how much carbon was emitted to produce the goods and services that people in that country actually consume? This "carbon footprint" includes the emissions from making a smartphone in one nation, shipping it to another, and selling it to a consumer in a third. To answer this, researchers use complex accounting tools called multi-regional input-output tables. Think of these as massive, interconnected spreadsheets that track how money and materials flow between every industry and every country in the world. By following these flows, experts can trace the hidden emissions embedded in our daily purchases. These tools have become essential for climate policy, helping governments and companies see the full environmental cost of their economies. However, for these numbers to guide real-world decisions, they must be trustworthy and align with the official records kept by national governments.

A recent study by researchers from several European institutions investigates a critical flaw in how these carbon footprints are currently calculated. The team compared three popular academic databases—EXIOBASE, GLORIA, and FIGARO-E3—against the official European Union database known as FIGARO. While the academic versions are praised for their high level of detail, breaking the economy down into hundreds of specific industries, the researchers wanted to know if this detail came at the cost of accuracy. They treated the official EU database as the gold standard, the benchmark against which all others should be measured, because it is built directly from the verified statistics that governments submit to international bodies. The study focused on the year 2015, as this is the only year for which the high-resolution FIGARO-E3 database is currently available for a direct comparison.

The researchers found that the high-resolution academic databases often tell a very different story than the official statistics. When they compared the monetary values in the tables—the actual numbers representing how much money flows between industries—they discovered significant mismatches. For the EXIOBASE and GLORIA databases, the differences were substantial. In some European countries, the total carbon footprint calculated using these academic tools differed from the official figure by as much as 25 to 44 percent. The researchers call this the "official statistics consistency bias." It turns out that the very effort to make these databases more detailed and granular sometimes introduces errors that make the final numbers drift far from reality. In many cases, this error caused by inconsistency with official data was actually larger than the error caused by grouping industries together too broadly, which is the problem these detailed databases were originally designed to fix.

The study also looked at specific products to see where the discrepancies were most severe. For items like machinery, electricity, and food products, the differences between the academic estimates and the official numbers were often even more pronounced than the national totals. In some instances, the academic databases suggested a product had a much lower carbon footprint than it actually did, while in other cases, they suggested it was much higher. By contrast, a newer database called FIGARO-E3, which was specifically built to combine high detail with strict adherence to official records, showed almost no deviation from the benchmark. Its results were nearly identical to the official figures, proving that it is possible to have both high detail and high accuracy, provided the data is anchored firmly to official sources.

These findings suggest a difficult trade-off for policymakers and researchers. For years, the scientific community has pushed for more detailed data to avoid the "sectoral aggregation bias," which occurs when diverse industries are lumped together, masking their true environmental impacts. The study confirms that while high detail is valuable, it is useless if the underlying numbers do not match the official records. If a government uses a database that deviates by 40 percent from official statistics, any policy based on that data could be fundamentally flawed. The authors argue that for carbon footprints to be used in serious climate policy, they must be consistent with the official data that governments already trust. While the detailed academic databases remain powerful tools for research and for exploring global supply chains where official data is scarce, they should be used with a clear understanding of their potential to mislead. The path forward, the study suggests, lies in creating new, high-resolution databases that are built directly on top of official statistics, ensuring that the numbers guiding our climate future are both detailed and undeniably accurate.