The Weaponization Objection: Why Public Welfare Scores Cannot Be Reliably Captured by Short-Sellers or Activists

Weaponization works against quarterly opaque scores, not Decision Accounting

The paper accepts the short-seller objection under naive implementation and rejects it under Decision Accounting. The difference is institutional design: latency, opacity, missing counterfactual records, and weak verification create the attack surface.

• Naive β_W: periodic publication, proprietary or undocumented scoring, no counterfactual record, limited pre-publication verification.
• Decision Accounting β_W: continuous updates, documented scoring function, Field 17 counterfactual records, ex ante signal verification.
• The objection targets an implementation the paper says no serious β_W regime should use.

The short-seller needs a fabricated signal to move β_W before correction

The weaponization strategy is modeled as a trading game around a firm F, true welfare state W* ∈ [0,1], public score β_W ∈ [0,1], administrator A, and short-seller S.

• Period 0: S observes private signal s_p and chooses whether to create fabricated signal s_f at cost c_f.
• Period 1: A observes S ∪ s_f and publishes β_W1.
• Period 3: S closes the short position and earns π = (P₀ - P₁) × Q before the score is corrected at Period 4.

Weaponization pays only when score impact beats fabrication cost and expected penalties

The paper’s profit condition is E[π] = δ × λ × Q - c_f - E. The strategy is profitable only if δ × λ × Q > c_f + E.

• δ is the score decline induced by the fabricated signal.
• λ is price impact per unit change in β_W.
• τ is the lag between manipulated publication and correction; quarterly publication makes τ large, continuous updates shrink it.

Quarterly opaque publication gives fabricated inputs time to affect price

Under the naive regime, the administrator can incorporate a fake welfare signal before detecting it, and the trader can exit before the correction appears.

• Attack examples in the paper include a fake whistleblower report, manipulated environmental audit, or fabricated customer complaint.
• No intra-period correction means the manipulated β_W can persist until the next fixed publication cycle.
• LIBOR, FX fixes, ISDAFIX, and credit ratings are named as benchmark-manipulation precedents.

Verification, audit trails, and continuous updates remove the trading window

Decision Accounting blocks the strategy by changing when signals enter the score, how they are checked, and how corrections are made visible.

• Unverified signals are flagged, assigned low weight, or excluded before they affect β_W.
• Continuous publication means correction occurs within hours or days rather than months.
• Field 17 records received signals, verification actions, scoring decisions, and counterfactual scores under alternative signal sets.

The attack relies on private knowledge that the administrator cannot test in time

The fabricated signal works only if the short-seller knows both the signal and its falsehood while the administrator and market do not. Decision Accounting narrows that gap.

• Signal verification cross-references whistleblower claims, company records, regulatory filings, third-party audits, and independent data sources.
• Pattern detection adjusts verification when a signal type repeatedly causes negative β_W movements that later reverse.
• Market monitoring uses the counterfactual record to detect traders who repeatedly profit ahead of β_W corrections.

Naive β_W turns a governance metric into a manipulation target

The bad equilibrium is a Hollow Win: the short-seller gains privately while destroying the public value of an informative welfare score.

• Score degradation: β_W becomes less informative about W*, so market participants discount it.
• Crowding out: fabricated inputs make genuine signals from whistleblowers, auditors, and regulators harder to separate.
• Regulatory response: rigid rules may reduce score flexibility and timeliness, making informativeness worse.

Decision Accounting makes suspicious β_W movements less profitable over time

Under Decision Accounting, manipulation does not occur in equilibrium because expected weaponization profit is non-positive for feasible parameter values.

• Verification improves as detected gaming attempts update the protocol, following the DMSE framework.
• Markets learn from the full history of β_W scores, corrections, and counterfactual records.
• λ falls for suspicious score movements because traders wait for confirmation before repricing.

The formal result separates β_W from the naive regime

Proposition 1 says weaponization is profitable under N iff δ × λ × Q > c_f + E[penalty_N], while π(S, DA) ≤ 0 for all feasible parameter values.

• Feasibility under DA requires that fabricated signals cannot pass verification with probability greater than ε, where ε → 0 as verification matures.
• Proposition 2 says E_N has score degradation, rising verification costs, signal crowding, and regulatory rigidity.
• Proposition 4 says the no-manipulation outcome sits on the Private Pareto frontier under Decision Accounting.

A successful 5% abnormal-return attack would falsify the theory

The paper gives an empirical failure condition: proper Decision Accounting is falsified if a short-seller weaponizes β_W and earns abnormal returns exceeding 5% over a 90-day window.

• The jurisdiction must use at least daily β_W updates, transparent construction, Field 17 records, and ex ante verification.
• The trader must generate or amplify a fabricated welfare signal, close within 90 days, and exceed 5% abnormal returns.
• must link fabricated signal to β_W change, β_W change to price movement, and the trade to premeditated strategy.

Legitimate negative research is price discovery, not weaponization

The paper separates fabricated welfare information from real negative information. Decision Accounting should admit the latter and exclude the former.

• A short-seller who finds a genuine welfare problem performs price discovery when β_W falls after verification.
• Adaptive gaming is answered by adaptive verification, not by abandoning β_W.
• Verification cost is weighed against the value of accurate welfare measurement for systemically important firms, firms with major externalities, and firms receiving public subsidies.

Benchmark design must target latency, opacity, missing records, and weak verification

The paper’s policy implication is narrow: do not implement β_W as a static periodic score. Build it as market infrastructure with auditable inputs and falsifiable safeguards.

• Continuous publication reduces τ, limiting exit-before-correction trades.
• Transparent scoring reduces the short-seller’s informational edge about how a signal will move β_W.
• Counterfactual records and penalties raise E by making manipulation reconstructable and prosecutable.