What trading volume measures, and when it misleads

Data providers build market-wide volume by collecting figures from individual venues and adding them together. The aggregation introduces several choices that are rarely visible to the reader. Which venues are included, and on what basis a venue qualifies for inclusion, is a judgement rather than a measurement. Whether derivative volume is counted alongside spot volume changes the total substantially, because futures and perpetual contracts trade far more than spot on most venues. And whether a venue's figures are adjusted for wash trading, or excluded entirely, depends on the provider's methodology.

The result is that two reputable providers can publish market-wide volume figures that differ by a large factor on the same day, and both can defend their number. A reader comparing volume across sources should treat the figures as indicative of relative activity rather than as measurements of an absolute quantity. The direction of change is usually more reliable than the level.

Exchange-reported volume has a well-documented credibility problem. Because venues compete for listings, visibility and the appearance of liquidity, there is a commercial incentive to report activity that did not occur. The practice is called wash trading: a venue, or a participant acting with its knowledge, buys and sells to itself to generate the appearance of volume. The trades are real in the sense that they settle, and they are not real in the sense that any economic risk changed hands.

Wash trading is difficult to detect from outside a venue because the trades look like ordinary trades in the public record. Researchers have used statistical fingerprints to identify it — patterns in trade sizes, timing regularities, and the relationship between volume and the number of active addresses — and those studies have found that reported volume on some venues has substantially exceeded plausible genuine activity. The finding is not that all volume is fake. It is that the figure cannot be taken at face value, and that the venues with the largest reported numbers have not always been the ones with the most genuine activity.

The practical response for a reader is to treat volume as a weak signal and to prefer measures that are harder to fabricate. Order book depth is more informative than volume because it represents a commitment to trade at a stated price, though it too can be manipulated by orders that are never filled. The number of distinct participants, where it is available, is more informative still. The liquidity page sets out why volume and liquidity are different properties, and the exchange price differences page explains how the gaps between venues relate to the depth actually available on each.