BIS paper: Bitcoin transfer estimates vary up to sixfold

A Bank for International Settlements working paper published Sept. 15 found that estimates of value transferred on Bitcoin can vary by as much as sixfold depending on how analysts translate raw ledger records into economic transfers.

The paper tested multiple measurement approaches on public blockchains and showed that choices about grouping and interpreting recorded movements produce sharply different totals for value transferred on Bitcoin. The sixfold figure describes the range across methods tested, not a universal error factor for every estimate.

For Bitcoin, the paper highlights transaction aggregation as a major source of divergence. Calculations depend on rules that group individual on-chain movements into higher-level transfers. A chart or headline number can be internally consistent while relying on assumptions about which ledger entries count as comparable economic events.

The analysis extends to programmable chains. The authors classified about 13 million active smart contracts, including roughly 1.4 million tokens, and found rapid contract and token issuance makes it harder to identify economically meaningful activity from raw counts. The paper reports that trading on many networks is highly concentrated and often centered on stablecoins rather than native tokens.

Stablecoin activity varied by network. The same stablecoin reflected different patterns depending on the blockchain. On Ethereum, stablecoin flows were more often linked to smart-contract interactions. On Tron, stablecoins were more frequently held outside contracts in patterns described as consistent with transactional use and store-of-value behavior. Treating every recorded unit or transfer as economically equivalent can blur distinct forms of activity and distort cross-chain comparisons, the paper notes.

Given these findings, the paper recommends treating on-chain indicators as approximations rather than direct measures of economic activity. It proposes producing granular, data-bounded estimates that make methodological assumptions explicit and using technical classification and disaggregation to connect ledger events to economic meaning. The paper advises researchers to disclose counting rules so users of on-chain metrics can judge what the figures represent.

The working paper frames its conclusions as the views of its authors and not necessarily official BIS positions.