Question 1

How is data quality controlled? (outlier removal and smoothing)

Accepted Answer

The commercial factors are cleaned (i.e., not raw multipliers). Four steps, in this order: Non-positive values and sectors with tiny output are dropped. Statistical filters remove implausible values. Gaps are filled from the country's own figure where it survived, otherwise its region group's average, otherwise the world average. The series is smoothed to damp the noise in input-output data. Screening comes before the averages are formed, so one extreme cell cannot distort the average other countries then receive. The most recent year is left unsmoothed, so it matches the underlying calculation. The Standard factors carry a quality flag recording all of this (see below). The cleaning is needed because EXIOBASE was built for national and global totals, which are constrained to official statistics and are reliable. An individual country-and-sector cell never was, because no aggregate result depended on it, and the accuracy of the total does not transfer to the cell. The filters are calibrated on greenhouse-gas intensities. Water, land use, ecosystem quality and human health have intensity distributions orders of magnitude apart, so the same limits would remove figures on the wrong basis. Until per-category limits are set, those categories receive only the checks that hold regardless of scale: impossible values, and sectors too small to form a reliable ratio.

Question 2

How do I know how reliable an individual factor is? (quality flags)

Accepted Answer

Every factor workbook carries a flag sheet mirroring the factor sheets cell for cell, so you can see what happened to any figure before you use it. In releases up to and including v3.11.1 the sheet is called "Outlier info" and the flag is a true/false value: whether the value was flagged as an outlier and replaced by a regional or world average. The current generation of workbooks renames the sheet to "Data Quality Flags" and reports one of four states, so that a partially and a wholly substituted cell are no longer reported identically: Flag What it means How to treat it Original The country's own figure, unchanged by the quality steps Use directly Time-smoothed The country's own figure, adjusted by the smoothing across years Use directly Regional average The country's own figure did not survive the checks; this is its region group's average, or the world average where the group has none Usable; disclose if the figure is material Missing No usable figure at any level; the cell is blank Do not use Figures aggregated from many cells (market-average factors, or factors on a standard sector classification) report the mix instead of a single label: Original, Mostly original, Mostly replaced, Replaced, or Missing. "Regional average" marks a figure that is not specific to that country, rather than one that is wrong. For a small country with a thin underlying sample, a regional average is frequently the better estimate. Two countries both carrying a regional average carry the same number, so country-level comparisons within a region should be made carefully. Either way you can see, and defend to an auditor, where each number came from.

Question 3

Are there uncertainty estimates?

Accepted Answer

No true uncertainty can be computed: like all global MRIOs, EXIOBASE is built from national statistics, trade, energy and production data, none of which publish uncertainties. Assumed-range error propagation is possible but produces pseudo-uncertainties that can give false confidence, so we don't ship them by default. For a specific sensitivity analysis, get in touch.

Question 4

Is EXIOBASE accepted for regulatory reporting?

Accepted Answer

Spend-based EXIOBASE factors are the GHG Protocol's recommended source for Purchased Goods & Services (Category 3.1) and a fallback elsewhere, and the PCAF Standard names them for financed emissions; SBTi asks for FLAG emissions to be reported separately, which the dataset supports. For specific regimes (e.g. CSRD/ESRS), confirm the current requirement. These build on the GHG Protocol, but XIO does not certify regulatory acceptance on your behalf.

Question 5

How accurate are the factors, and when are they not the right tool?

Accepted Answer

Spend-based factors are screening-grade: strong for covering an entire footprint consistently and finding hotspots, less precise for any single line item, because each reflects a sector-and-region average rather than a specific product or supplier. They suit Scope 3 screening, financed emissions and gap-filling. They are the wrong tool where you need product-specific precision, where a supplier's own verified data exists (use that), and for the downstream categories they do not cover. Use them where their breadth is the strength, and pair them with activity data where precision matters.

Question 6

Can I use the factors for CBAM or a specific regulation?

Accepted Answer

Spend-based factors are a screening and gap-filling tool aligned with the GHG Protocol and PCAF (above). They are not a substitute for the specific embedded-emissions data some regulations require. CBAM, for example, is built around installation-level embedded emissions of imported goods, a different measurement from a spend-based estimate. Use EXIOBASE where screening or fallback estimates are accepted, and check the specific regime's rules for what it will and won't accept.

Questions, answered

Current release at a glance

Getting Started

Data Coverage & Structure

Choosing the Right Emission Factor

Data Quality & Reliability

Data Quality & Reliability