How to read welfare beta (βW).

Welfare beta (β_W) = annual system-welfare loss (ΔW) / annual industry revenue (Π). β_W is pronounced beta W. ΔW is pronounced delta W. Π is pronounced capital pi. The denominator is revenue, never profit. ΔW and Π must use the same domain, same time period, and same activity boundary.

Domain beta-W values remain withheld until the row has an admitted result. Its interval is the fifth-to-ninety-fifth percentile range from the declared simulation model; the result must carry source-backed channel inputs, an independently sourced denominator, matched activity boundary, dependence receipt, seed, draw count, packet/result hashes, and an independent rerun receipt.

The portfolio aggregate remains withheld pending evidence-backed re-estimation. A public aggregate interval requires a signed cross-domain overlap ledger, dependence model, and rerun receipt; those gates remain open.

ΔW is an accounting attribution of annual system-welfare loss to a domain on a matched revenue boundary, not a causal-identification estimate. The numerator and denominator must cover the same domain, same time period, and same activity boundary; that discipline makes the row reconstructable, but it does not by itself prove that the revenue caused every dollar of loss.

Where the research program makes a causal or asset-pricing claim, it is stated separately and tested separately. The return-side companion paper, for example, uses a within-firm, firm-fixed-effects event study and reports the result narrowly: βW becomes return-relevant at legibility events, while the broad factor interpretation is rejected.

Cite Domain Tables for canonical βW, Π, ΔW, Monte Carlo interval status, row status, and evidence routes. Treat public-paper prose, decks, and summaries as interpretive aids unless they route back to the table or study record. Do not create new βW values by swapping profit, margin, market capitalization, or transaction volume into Π.