Decision Accounting as a Report-Incentive Mechanism

Decision accounting turns the missing welfare coordinate into a reportable state

The sixteenth field of a Decision Accounting record carries system-welfare impact W. This paper models truthful reporting as an incentive-compatible mechanism, not a slogan.

• Decision-maker bears cost κ to truthfully record W
• Misreport detected ex post with probability p by a Conflictoring auditor
• On detection, decision-maker bears liability L

Truthful recording is a best response iff pL ≥ κ

Proposition D.1: Misreporting saves κ and risks pL; truthful recording pays κ with no sanction. Truthful recording is weakly preferred when expected sanction covers recording cost.

• p = detection probability (endogenized from mover audit equilibrium)
• L = liability (endogenized via external legislative mover)
• κ = cost of assembling and disclosing the system-welfare record

p is the equilibrium audit rate of the mover set

Each mover m audits iff its expected recovery exceeds audit cost: θ β_m ≥ α . The decision-maker faces detection probability p = 1 - ∏ (1 - θ ) over auditing movers.

• α = audit cost for mover m
• θ = probability audit succeeds in establishing misreport
• β_m = private return to mover (verdict, mandate, repriced position)

The writer cannot foresee which movers will audit

The unpredictable-reader property (Prat, 2005) makes p exogenous to the writer's strategy. The writer best-responds against the set of potential auditors, not a chosen one.

• Writer cannot tune the record to a known auditor
• p is writer-independent, bounded above zero when one mover is un-foreclosable (Proposition C.1a)
• Closes the loop left open in the companion paper

L is endogenized through an external legislative mover

Liability is set by a mover outside the decision-maker's control, making it renegotiation-proof. The paper addresses four standard design objections.

• Determination of L: set by legislative mover
• Collectibility against judgment-proof agents: resolved via individual liability
• Participation constraint: must hold for decision-maker
• Verifiability of W: resolved via process-and-range verification

A truth-telling equilibrium exists under three conditions

Proposition D.2: When at least one mover is un-foreclosable, liability attaches to the named individual, and participation holds, a truth-telling equilibrium exists.

• Un-foreclosable mover ensures p > 0
• Individual liability prevents judgment-proof evasion
• Participation constraint ensures decision-maker engages

Four standard objections are resolved within the model

The paper defends the IC constraint against attacks on liability determination, collectibility, participation, and verifiability.

• L is not arbitrary: set by legislative mover
• Individual liability solves collectibility against judgment-proof agents
• Participation constraint is met when expected net benefit positive
• W is verifiable through process-and-range verification, not point accuracy

Decision accounting is a report-incentive mechanism, not a compliance checklist

The paper positions Decision Accounting above applied domain ranking as theory. It is about report incentives, not merely a compliance checklist.

• Nearest literature: Prat, Becker, Myerson-Satterthwaite, mechanism design, audit and liability
• Field 17 carries system impact; Field 16 makes predictions scoreable
• Record must survive readers the writer cannot fully predict

A record is not a reform by itself

Decision Accounting only changes behavior when records are expected to be read, scored, and tied to consequences. Without pL, the form degenerates into paperwork.

• No empirical β_W tile in this paper
• Draft theory statement; peer review pending
• Classroom use: vary p, L, and κ to show when truth-telling is cheaper than concealment

The missing coordinate becomes a reportable state with teeth

The paper turns the Hollow Win—where parties transact while the system fails—into a game where the system's interest is carried into decision-making via incentive-compatible reporting.

• Changes the game from one where system welfare is omitted to one where it is reported
• Truthful recording is a best response when pL ≥ κ
• Opens the door to applied β_W domain ranking in future work