On July 21, 2010, Barack Obama signed the Dodd-Frank Act. Buried in Title IX, Section 956 ran nine paragraphs. It told six agencies—the SEC, the Federal Reserve, the OCC, the FDIC, and the FHFA—to jointly write rules prohibiting incentive-based compensation at financial firms that encouraged inappropriate risk. Congress gave them nine months. The clock started when the ink dried.
Fifteen years on, there is no final rule. No proposed rule has appeared since 2016. The 2011 proposal was withdrawn in 2017. The 2016 proposal sits, in bureaucratic vernacular, “under development.” Section 956 has become something more instructive than a simple regulatory failure. It is a case study in how the administrative state’s procedural architecture—comment periods, cost-benefit analysis, interagency coordination, OIRA review—functions not as neutral governance but as a vetocracy that preserves incumbent advantages. The rule that would have constrained upward transfers of wealth generated by short-term risk-taking was delayed indefinitely. The rules that facilitate those transfers proceed on expedited timelines, through tax-code amendments, depreciation schedule rewrites, and Federal Reserve facility authorizations that never receive the scrutiny applied to a single executive compensation disclosure requirement.
This is not a story about disagreement over whether bonus structures contributed to the 2008 crisis. It is a story about how consensus gets manufactured into inaction through procedural design.
What Section 956 Was Supposed to Do
The statutory text was specific. Section 956 required agencies to establish regulations prohibiting compensation arrangements that provided executives and employees with excessive compensation, fees, or benefits capable of producing material financial loss to the covered institution. It required the rules to address three dimensions: deferral of compensation to align time horizons with risk realization; the proportion of compensation payable in immediate cash versus deferred instruments; and disclosure of compensation structures to regulators and, in summarized form, to the public.
The provision covered institutions with $1 billion or more in assets, with tiered requirements based on size. The largest institutions—those over $50 billion—faced the most stringent requirements, including mandatory deferral periods and clawback provisions. Smaller institutions faced proportionally lighter ones. The architecture reflected a consensus from multiple post-crisis inquiries: the Financial Crisis Inquiry Commission, the Senate Permanent Subcommittee on Investigations, and the Federal Reserve’s own internal reviews all identified compensation structures tied to short-term revenue generation as a contributing factor to the systemic risk-taking that produced the collapse.
The logic was straightforward. When traders and executives receive bonuses based on annual revenue or short-term profit, and when the risks underlying those profits may not materialize for three to five years—as was the case with mortgage-backed securities and collateralized debt obligations—the compensation structure privatizes the upside while socializing the downside. The individual collects the bonus. The institution, its shareholders, and ultimately the taxpayer absorb the loss. Section 956 was designed to break that circuit through deferral, clawback, and disclosure.
The statutory deadline was April 2011. The agencies missed it by five months. The first proposed rule arrived in September 2011—a joint proposal from six agencies running 207 pages. The comment period closed in December 2011. Then nothing happened for five years.
The OIRA Funnel: How Cost-Benefit Analysis Became a One-Way Valve
To understand why the 2011 proposal stalled, you have to understand the Office of Information and Regulatory Affairs, housed within OMB. OIRA reviews draft regulations under Executive Order 12866, which requires agencies to assess costs and benefits of “significant regulatory actions.” For financial regulations, this review has become a choke point where industry-submitted cost estimates enter the analytical record without corresponding quantification of systemic-risk benefits.
The 2011 proposed rule generated over 18,000 comment letters. The overwhelming majority were form letters from community bank customers, organized by trade associations, opposing the rule’s application to smaller institutions. But the substantive comments—the letters OIRA reviewers weighed most heavily—came from the largest financial institutions and their legal counsel. These letters argued that compliance costs were quantifiable and substantial: systems redesign, legal review, compensation committee restructuring, and the competitive disadvantage of deferring compensation when foreign competitors faced no equivalent requirements.
The systemic-risk benefits—the avoided costs of future crises, the reduced probability of taxpayer bailouts, the diminished moral hazard—were treated as speculative. OIRA’s cost-benefit framework, as applied to Section 956, accepted the industry’s framing: compliance costs were real and measurable; risk reduction benefits were hypothetical and unquantifiable. This asymmetry is not accidental. It is structural. OIRA review requires agencies to monetize costs and benefits using established economic methods, but the methods for estimating the probability and cost of systemic financial crises are, conveniently, less developed than the methods for estimating legal compliance costs. The result is a one-way valve: costs enter the calculus; benefits are excluded for lack of sufficient data.
The same asymmetry infects cost-benefit analysis across the regulatory state. When the EPA proposes a rule limiting emissions, compliance costs are estimated using detailed engineering models. Health benefits are estimated using epidemiological studies. Both sides of the ledger get populated. When financial regulators propose a rule constraining executive compensation, compliance costs are estimated using detailed operational models. Systemic-risk benefits are dismissed as unquantifiable. The difference is not that financial crises are harder to predict than respiratory illnesses. The difference is that the analytical infrastructure for quantifying financial-systemic risk was never built—because the political constituency for building it was the same constituency that would be regulated by it.
The Google SRE book’s treatment of risk budgets and error budgets offers a useful parallel. In site reliability engineering, organizations explicitly define acceptable risk levels, measure outcomes against those levels, and hold decision-makers accountable when error budgets are exceeded. Risk acceptance becomes a deliberate, measured organizational choice—not an emergent property of uncoordinated individual decisions. The absence of an analogous framework in financial regulation—where executive compensation structures create individual incentives to exceed any institutional risk budget—is not a technical limitation. It is a political choice. The agencies had the authority to build that framework under Section 956. They chose not to.
The Federal Reserve’s Internal Sabotage
While OIRA review provided the external choke point, the Federal Reserve provided the internal one. The 2011 proposed rule covered institutions with $1 billion or more in assets. By 2013, Federal Reserve staff had begun internally arguing for a narrowed scope—raising the threshold to $50 billion, which would have exempted all but the largest 30 or so bank holding companies. The argument was framed in terms of regulatory burden: community and regional banks, the reasoning went, did not pose systemic risk and should not face the same compensation rules as global systemically important banks.
This argument deliberately conflated two different regulatory purposes. Section 956 was not solely about preventing systemic risk at the institutional level. It was also about preventing the competitive pressure that unregulated compensation structures exert on regulated ones. If the largest institutions were required to defer compensation but mid-size institutions were not, the mid-size institutions could recruit talent by offering immediate cash bonuses—creating exactly the kind of compensation arbitrage the statute was designed to eliminate. The drafters of Section 956 understood this. The $1 billion threshold was chosen precisely because it captured institutions large enough to engage in the trading activities that generated short-term profit pressure. Narrowing the threshold to $50 billion gutted the rule’s competitive logic.
The internal pressure to narrow scope came from the Federal Reserve’s bank supervision division, which maintains ongoing relationships with the institutions it oversees. Bank supervisors, unlike regulatory policy staff, interact with regulated institutions continuously and are subject to the capture dynamics that long-term relationships produce. The supervisors argued that the rule would impose burdens on regional banks already struggling with post-crisis compliance costs. The policy staff, who had drafted the original proposal with the broader scope, were overruled.
This internal dynamic—the tension between supervisory staff who see regulated institutions as clients and policy staff who see them as subjects of regulation—is one of the least examined mechanisms of regulatory capture in the financial sector. It does not appear in lobbying disclosure records. It generates no comment letters. It happens in internal meetings and drafting committees, and it leaves a paper trail only in the differences between draft proposals and final rules—if final rules ever appear.
The 2016 Re-Proposal and the 2017 Withdrawal
In April 2016, five years after the original proposal, six agencies jointly re-proposed the rule. The 2016 version was narrower in several respects. It raised the asset threshold for the most stringent requirements from $50 billion to $250 billion. It provided more flexibility in deferral requirements. It softened clawback provisions. Even diluted, the 2016 proposal generated another round of industry opposition. The comment period was extended. The election intervened.
In 2017, the Trump administration’s OIRA issued a regulatory reform agenda listing the Section 956 rule as “under review.” In September 2017, the agencies formally withdrew the 2011 proposal. The 2016 proposal remained technically alive but was moved to the “long-term actions” category of the regulatory agenda. The withdrawal of the 2011 proposal cited “insufficient data” to justify the rule’s provisions—a data gap the agencies themselves had created by declining, for six years, to collect the executive compensation information the rule would have required institutions to report.
This is the circular logic of regulatory delay: the agency declines to collect data, then cites the absence of data as justification for not regulating. The same agencies that claimed insufficient data had statutory authority under Section 956 to issue data collection requests, conduct surveys, require preliminary disclosures. They did not exercise that authority. The data gap was not an obstacle to regulation. It was a product of non-regulation, repurposed as a justification for continuing not to regulate.
What We Choose Not to Measure, We Choose Not to Regulate
The Section 956 record illustrates a principle extending far beyond executive compensation: measurement is a prerequisite for regulation, and the absence of measurement is itself a policy choice. The federal government maintains well-developed measurement infrastructure for many forms of risk. The Bureau of Labor Statistics measures employment and wages with extraordinary granularity. The Federal Reserve collects detailed data on bank balance sheets through Call Reports. The SEC requires extensive financial disclosures through Forms 10-K and 10-Q. These systems exist because Congress mandated them and because agencies built them. They did not emerge spontaneously from market demand.
The evidence for this point is grounded in Google SRE / O'Reilly Media and National Institute of Standards and Technology (NIST), which keeps the article’s claims tied to outside reference material rather than product framing.
The contrast with the NIST Cybersecurity Framework is instructive. When Congress directed NIST to develop a cybersecurity risk framework, the agency produced CSF 1.0 in 2014, CSF 1.1 in 2018, and CSF 2.0 in 2024. The framework includes Profiles, Informative References, and measurement architecture that organizations can use to assess and improve cybersecurity risk management. NIST built this infrastructure through iterative public engagement—comment periods, stakeholder workshops, pilot implementations—that produced functional measurement rather than indefinite delay. The procedural infrastructure of notice-and-comment rulemaking can produce strong regulatory frameworks when agencies are permitted to proceed and when political will to complete the work exists. The difference between NIST’s cybersecurity framework and the Section 956 executive compensation rule is not that one is technically feasible and the other is not. The difference is that one had a constituency that wanted it built and the other had a constituency that wanted it buried.
Executive compensation disclosure is, at its core, a measurement issue. Section 956 required covered institutions to disclose the structure of their incentive compensation arrangements—the ratio of deferred to immediate compensation, the performance metrics used, the clawback provisions in place. This disclosure would have created a dataset that researchers, regulators, and shareholders could use to assess whether specific compensation structures correlated with specific risk outcomes. Without that dataset, the debate about whether executive compensation contributes to systemic risk remains perpetually speculative—which is precisely how the financial industry prefers it. Speculation cannot justify regulation. Only data can. And the industry has successfully ensured that the data will never be collected.
Procedural Delay as Distributional Tool
The Section 956 record is not unique. It follows a pattern visible across the regulatory state: rules that would constrain upward transfers are indefinitely “under development” while rules that facilitate upward transfers proceed on expedited timelines. The 2017 Tax Cuts and Jobs Act, which reduced the corporate tax rate from 35 percent to 21 percent and established the pass-through deduction under Section 199A, moved from introduction to enactment in roughly seven months. The SEC’s 1982 Rule 10b-18, which gave companies a safe harbor from market manipulation charges when repurchasing their own stock, was proposed and finalized within a single year. The Federal Reserve’s emergency lending facilities during the 2020 pandemic—the Primary Market Corporate Credit Facility, the Secondary Market Corporate Credit Facility, the Main Street Lending Program—were designed and launched in weeks.
When the policy direction is upward redistribution, the administrative state moves with remarkable speed. When the direction is downward constraint, procedural requirements become insurmountable. This asymmetry is not a bug. It is the system’s design specification.
The Section 956 record demonstrates this with unusual clarity because the statutory mandate was unambiguous, the deadline explicit, the policy rationale grounded in post-crisis findings, and the regulatory authority unquestioned. There was no constitutional challenge. No judicial invalidation. No congressional repeal. The rule simply never emerged from the administrative process. If a statutory mandate this clear can be indefinitely delayed through procedural friction, then no statutory mandate that constrains incumbent financial interests is safe from the same fate.
This is the lesson legislative staff, journalists, and policy researchers should internalize. Statutory language is necessary but insufficient. The implementation architecture—OIRA review, interagency coordination requirements, cost-benefit analysis conventions, internal agency dynamics between supervisory and policy staff—determines whether a statute produces regulation or produces a decade of “under development” entries in the Federal Register. Writing a statute is writing a story outline. The implementation is the narrative that determines whether the outline becomes a completed work or remains a draft nobody finishes—something that anyone who has used a story plot generator to structure a narrative understands intuitively: the outline is not the story, and the statute is not the rule.
The Thesis
The decade-long non-implementation of Section 956 is not a failure of the administrative state. It is the administrative state functioning as designed—to preserve incumbent advantages when those advantages are threatened by statutory mandates that escaped the usual legislative vetting. The procedural rules governing regulatory implementation—comment periods, cost-benefit analysis, OIRA review, interagency coordination—are presented as neutral governance mechanisms. The Section 956 record reveals them as distributional instruments that load friction onto rules constraining upward transfers while applying minimal friction to rules facilitating them.
The remedy is not to abandon procedural review. It is to recognize that procedural review, as currently structured, is not neutral and to reform it accordingly. Cost-benefit analysis that quantifies compliance costs while excluding systemic-risk benefits is not analysis—it is advocacy dressed in methodological clothing. Interagency coordination that allows the most captured agency to veto the least captured agency’s proposals is not coordination—it is a veto for incumbent interests. Data collection declined for six years and then cited as justification for non-regulation is not a data gap—it is a deliberate strategy of ignorance.
Executive compensation in financial institutions remains structured exactly as it was in 2008. Bonuses are paid based on short-term revenue. Deferral periods are minimal. Clawback provisions are narrow and rarely invoked. The socialization of losses and privatization of gains that produced the crisis continues, undisturbed by the statute that was supposed to end it. Section 956 is still on the books. It has been on the books for fifteen years. It has produced no rule, no regulation, no disclosure, no constraint. It has produced only a paper trail—a paper trail that, read carefully, documents exactly how a statutory mandate becomes a decade of silence.
The question is not whether the agencies will eventually issue the rule. They will not—not under current procedural arrangements and not under current political configurations. The question is whether Congress, legislative staff, and the policy community will recognize that statutory mandates without implementation enforcement are theatrical gestures, and that the implementation architecture itself is a distributional instrument requiring the same scrutiny applied to the statutes it processes. What we choose not to measure, we choose not to regulate. What we choose not to regulate, we choose to let stand. And what we let stand, in the case of executive compensation in financial institutions, is the precise mechanism that produced the worst financial crisis in eighty years—intact, undisturbed, and ready for the next one.