Brokerage Against the Public

Brokered regulatory games pay the broker to keep the public authority at a distance

The paper models a regulated party A, a broker B, and a public authority or system C. When B's revenue rises with the information asymmetry δ(A,C), the stable outcome is Hollow Win: C=0, A=1, B=1.

• Apple's Double Irish with a Dutch Sandwich routed profits through Irish entities tax-resident nowhere, a Dutch shell, and a Bermuda mailbox.
• Apple paid a 0.005% effective tax rate on European profits in 2014.
• The European Commission ordered €13 billion in back taxes to Ireland in 2016.

The bilateral payoff record cannot see the tax base, financial system, or environment

The Missing System Theorem says W is absent from the payoff vector in any bilateral game G= A,B . Broker and client can record a private gain while the public system absorbs the loss.

• G has 4 outcomes because only A and B are scored.
• G1= A,B,C has 8 outcomes because C is scored too.
• The Hollow Win is (C=0,A=1,B=1); the Win-Win-Win is (1,1,1).

Exploit and maintain is the unique Nash equilibrium when W is excluded

A chooses exploit because brokered avoidance costs less than direct compliance. B chooses maintain because closing the hole reduces δ(A,C) and cuts B's revenue. No unilateral deviation improves private payoff.

• A1: π increases with δ(A,C), the cost of direct communication, verification, or coordination without B.
• A2: payoffs are computed as (π ,π ), with W excluded.
• A5: after t>t*, brokered bilateralism becomes protected by institutional entrenchment.

Disclosure adds facts without changing the payoff calculation

A disclosure rule expands I to I' = I ∪ w , but payoffs remain (π ,π ). If W is still not payoff-relevant, exploit and maintain remains an equilibrium.

• The paper calls this strategic documentation: ritual mention of system welfare without changing action.
• OECD BEPS and EU due diligence directives are treated as disclosure-style reforms that do not by themselves create C=1.
• Disclosure-only rules can raise β_W by increasing broker compliance revenue Π_B while dW persists.

Global tax advisory has β_W=5.3 on the paper's conservative estimate

The paper estimates tax advisory and structuring revenue Π at about 120 billion per year and facilitated corporate tax revenue loss dW at about 640 billion per year. β_W = 640B/120B = 5.3.

• β_W>5 is the paper's strong intractability range.
• The estimate excludes transfer-pricing manipulation of intangibles, treaty shopping, and hybrid mismatches outside BEPS estimates.
• It also excludes secondary losses from misallocated capital, degraded public services, and jurisdictional tax competition.

KPMG earned 124 million selling shelters that cut Treasury revenue by about 3.7 billion

From 1998 to 2005, KPMG's U.S. tax practice marketed shelter products with no economic substance. The IRS lacked real-time capacity to audit the complex transactions.

• Clients generated about 11 billion in phony tax losses and reduced tax liability by roughly 3.7 billion at a 35% marginal rate.
• KPMG earned about $124 million in fees, with estimated product profit margins of 40% to 50%.
• β_W ≈ 29.8, or about 30 of system welfare loss per 1 of KPMG revenue.

Confidentiality clauses aligned taxpayer and advisor against the IRS

The paper names this Conflictoring: formally adversarial parties share incentives against the public system. KPMG and clients under-disclosed the shelters' true nature while preserving the information gap with the IRS.

• Engagement letters included confidentiality clauses that blocked clients from sharing shelter structures with competitors or regulators.
• KPMG admitted criminal wrongdoing in 2005 and paid $456 million in penalties, deferred prosecution agreement fines, and restitution.
• The IRS still did not receive transaction-level data for real-time detection of similar shelters.

Issuer-paid ratings turned loan-quality opacity into $3.5 billion of rating revenue

From 2004 to 2007, Moody's and S&P sat between mortgage originators and investment banks on one side and institutional investors on the other. Their ratings converted granular loan risk into regulatory capital signals.

• Moody's and S&P earned about $3.5 billion from rating MBS and CDO products.
• Originators and issuers gained about $1.2 trillion in fees from originating and securitizing subprime mortgages.
• At the peak, structured finance was about 45% of Moody's revenue and 40% of S&P's revenue.

The paper maps inflated ratings to a 10 trillion to 15 trillion system loss

The rating agencies were paid by issuers, while investors and the financial system relied on their risk labels. The broker's revenue came from the side that benefited from high ratings.

• The IMF estimated $4.1 trillion in lost output for advanced economies from the crisis.
• The broader welfare loss, including unemployment, foreclosures, reduced retirement savings, and increased poverty, is estimated at 10 trillion to 15 trillion.
• The U.S. government spent $700 billion on TARP bailouts.

Field 17 makes system welfare auditable and penalty-relevant

Mandatory 17-field Decision Accounting changes the game from G= A,B to G1= A,B,E . Field 17, SYSTEM_WELFARE , must be populated with a verifiable method and visible to an enforcement authority E.

• E can impose penalties proportional to W when W<0.
• Exploit and maintain stops being an equilibrium when the penalty exceeds the private gain from exploitation.
• Field 17 must be independently verifiable, consistent across decisions, and subject to conflictoring review.

Estonia's X-Road reduces broker rents by creating direct bypass ties

X-Road is the paper's proof of concept for redundancy collapse. It creates k≥2 independent pathways between databases and citizens, reducing δ(A,C) to δ'(A,C)<δ(A,C)/k.

• With δ(A,C) near zero for routine transactions, brokers lose revenue from document retrieval, form filing, and inter-agency coordination.
• Estonia has 99% of public services online and a VAT gap below 2%.
• The paper says replication needs political commitment to eliminate broker rents, redundant infrastructure, and a legal mandate for Decision Accounting with Field 17.

Each theorem names observations that would break the model

The paper is testable because it specifies what would refute the Hollow Win, disclosure futility, Decision Accounting transformation, redundancy collapse, and β_W claims.

• F1: find a brokered regulatory game where Π_B rises with δ(A,C), W is excluded, and the stable outcome is not Hollow Win.
• F2: find a disclosure-only intervention that achieves C=1 without changing payoffs.
• F3: find implemented 17-field Decision Accounting with Field 17 that fails because records are unaudited or penalties are capped below private gains.

The remedy is game redesign, not more disclosure inside G

The paper's sequence is: identify the brokered bilateral disclosure game, name the Hollow Win, then apply Decision Accounting with Field 17 plus multi-audience access and enforcement.

• The Five-Minute Test asks whether the record makes the system welfare loss visible quickly enough for an enforcement audience.
• β_W diagnoses severity; β_W>5 means strong intractability in the paper's terms.
• The Reform Dividend is estimated at about $73.8T per year across brokered regulatory games.