The Most Misunderstood Banking Statute of Its Decade

On May 24, 2018, President Donald Trump signed the Economic Growth, Regulatory Relief, and Consumer Protection Act, Public Law 115-174, the most significant revision of the Dodd-Frank Act since its passage, and within hours two incompatible accounts of what he had just signed were in wide circulation. In one account, Congress had begun dismantling the post-crisis framework and the largest banks were the beneficiaries. In the other, Congress had done something far narrower: it had redrawn the lines that determined which institutions faced the toughest Dodd-Frank requirements, leaving the requirements themselves in place. The second account was the accurate one, but the first proved harder to dislodge, and the confusion between the two accounts shaped every later argument about the statute, including the fierce debate that followed the bank failures of March 2023.

This article is about the amendment stage of a statute’s life, and about a particular kind of amendment, one that changes coverage rather than substance. The 2018 statute did not rewrite what Dodd-Frank required of the financial system. It rewrote whom those requirements reached. Enhanced prudential standards, the Volcker Rule, stress testing, resolution planning, the derivatives clearing mandate, the Consumer Financial Protection Bureau, the Orderly Liquidation Authority, the Financial Stability Oversight Council: every one of these authorities survived the 2018 law intact. What changed was the set of institutions to which each authority applied, and the mechanism by which it applied, automatic in some cases, discretionary in others, unavailable in still others. To understand the statute is to understand that distinction, because nearly every misdescription of it, from the floor speeches of its opponents to the headlines about its aftermath, comes from collapsing the two.

The re-tiering, not repeal: the 2018 statute left every major Dodd-Frank authority intact and changed which institutions they reach, which means arguments about whether Dodd-Frank was repealed and arguments about whether the tailoring was calibrated correctly are entirely different arguments that are constantly conflated.

That claim is the organizing principle of everything that follows. It explains the vote count, which was bipartisan in a way financial legislation rarely was during that decade. It explains the structure of the law, which reads less like a repeal bill than like a new set of sorting instructions laid over the existing framework. And it explains the controversy that erupted five years later, when two large regional banks failed in March 2023 and the country discovered that the question of whether the tailoring had been calibrated correctly was a live and unresolved one, with serious people on both sides and a central bank review that refused to settle it cleanly.

The stakes of getting this right are not abstract. The Dodd-Frank Act was the most ambitious reorganization of American financial regulation since the Great Depression, and its 2018 amendment was the first large-scale test of whether Congress could adjust that architecture without breaking it. Readers who finish this article will be able to name what the 2018 statute actually changed, explain why sixteen Democrats and one independent joined every Republican present in the Senate to pass it, and evaluate honestly the contested question of whether raising the enhanced supervision threshold contributed to bank failures five years later. Those are the three competencies the brief for this article demands, and they are worth holding separately in mind, because each answers a different question and the public argument has a persistent habit of treating them as one.

What the Statute Is, and Where It Sits in a Statute’s Life

Every major federal statute has a life cycle. It is enacted in response to a crisis or a long-building problem, it is implemented through years of rulemaking and litigation, and then, if it survives, it enters the amendment stage, the long middle period in which Congress adjusts, trims, extends, and occasionally rewrites what it built. The amendment stage is where statutes are most often misunderstood, because amendments are technical, because they attract less press coverage than original enactments, and because their effects are usually conditional: this rule applies to fewer firms, that exemption covers more activity, this deadline runs longer. The 2018 Dodd-Frank rollback is a textbook specimen of the amendment stage, and it is an especially instructive one because the amendment it made was to coverage rather than substance.

The formal identity, stated once as the series standard requires: the Economic Growth, Regulatory Relief, and Consumer Protection Act, Public Law 115-174, enacted by the 115th Congress from Senate bill 2155 and signed on May 24, 2018. The name itself is a small study in legislative branding. Supporters called it regulatory relief and consumer protection; opponents called it a rollback and, in the sharpest formulations, a gift to Wall Street. The statute’s own text is less dramatic than either label. It runs to dozens of sections touching capital rules, mortgage origination, data reporting, examination schedules, custody arrangements, and consumer protections, but its center of gravity is unmistakable: five changes to the way Dodd-Frank’s requirements sort institutions by size and activity, surrounded by a larger number of targeted provisions for particular institution types.

To place the amendment in the statute’s life, it helps to recall what Dodd-Frank looked like at the start of the decade. The 2010 act had imposed its most demanding requirements through a single, simple sorting mechanism: total consolidated assets. A bank holding company with 50 billion dollars or more in total consolidated assets automatically faced enhanced prudential standards under section 165 of Dodd-Frank, including supervisory stress tests, risk committee requirements, liquidity standards, and resolution planning. The 50 billion dollar line was not the product of deep analysis about where systemic risk began; it was a legislative compromise, a round number chosen in the rush of post-crisis lawmaking. For the largest institutions, the ones whose failure could plausibly threaten the system, the line was uncontroversial. For the institutions just above it, regional banks with a few tens of billions in assets and traditional lending models, the line became, over the following years, the single most resented feature of the post-crisis framework. The 2018 statute was, at its core, Congress revisiting that compromise.

That is the series thesis for this article, and it is worth stating plainly: the amendment stage of a statute’s life is where legislatures decide whether the original law’s coverage still matches the problem it was written to solve, and the 2018 law is the case in which the amendment changed coverage rather than substance. Nothing in the sections that follow will make full sense without that frame. The threshold changes are coverage changes. The Volcker exclusion is a coverage change. The leverage ratio option is a coverage change, one that lets qualifying institutions choose a different way of measuring the same safety. Even the mortgage and reporting provisions are coverage changes, carving smaller lenders out of obligations written with larger ones in mind. And the parts of Dodd-Frank that the 2018 law left alone are the substance, the authorities themselves, untouched.

The Decade Between: From Crisis Thresholds to Complaints About Them

The Dodd-Frank Act was written in the shadow of the 2008 financial crisis, and its thresholds carry the marks of that origin. When Congress drafted the law in 2009 and 2010, the governing fear was that regulators had lacked both the authority and the information to see the next large failure coming, and the legislative answer was to make supervision automatic and comprehensive above a fixed asset line. The 50 billion dollar threshold for enhanced prudential standards was the most consequential of these lines. Cross it, and a bank holding company entered a regime of annual supervisory stress tests, company-run stress tests, mandatory risk committees, liquidity coverage obligations, single-counterparty credit limits, and resolution plans filed with the Federal Reserve and the Federal Deposit Insurance Corporation. Stay below it, and the company faced the ordinary supervisory regime that had existed before the crisis. The line was bright, simple, and easy to administer, which were its virtues. Its vice was that it treated a 55 billion dollar regional lender in the Midwest as a supervisory peer of a 2 trillion dollar money center bank, and the compliance costs of the enhanced regime did not scale down gracefully.

By the middle of the decade, the complaints had organized themselves into a political force. Community bankers, the leaders of the thousands of small institutions that held a small fraction of the system’s assets but originated a large share of its small business and agricultural loans, argued that Dodd-Frank’s fixed costs fell hardest on them. A stress testing regime designed for global trading houses was, in their telling, an absurd imposition on a bank whose balance sheet consisted mostly of local mortgages. The mortgage rules drew particular fire: the qualified mortgage standard and the expanded Home Mortgage Disclosure Act reporting obligations, both products of the post-crisis rulemakings, required data systems and legal review that small lenders could not spread across large volumes. Trade associations for community banks and credit unions made these arguments in hearing rooms for years, and they found a receptive audience among senators from states where community banks were major employers and civic institutions. The evidence on the underlying economic claim is the subject of the series companion on Dodd-Frank’s effects on the banking industry, which weighs the data on consolidation, lending, and compliance cost; what matters for this article is that the political claim, that the law’s burdens were miscalibrated for smaller institutions, had become by 2017 a bipartisan talking point rather than a partisan one.

The original thresholds, meanwhile, had been set by a Congress legislating under extreme time pressure in the aftermath of a crisis that had nearly broken the financial system, a history recorded in the series account of the crisis-era legislation of 2008. The lawmakers of 2018 were not legislating under that pressure. They were legislating in a long expansion, with bank profits at record levels, no failures of significant institutions in several years, and a widespread sense, at least outside the progressive wing of the Democratic Party, that the post-crisis framework had overshot in its treatment of smaller firms. That sense was the tailoring rationale, and it deserves a fair statement, because the later debate about the 2023 failures would repeatedly caricature it. The rationale was not that regulation was bad, or that the crisis could not recur. It was that regulatory intensity should vary with the risk an institution posed to the system, that a one-size threshold was a crude way to achieve that variation, and that the crudity had real costs in the form of consolidation pressure on small lenders and reduced credit availability in the communities they served. One could accept every element of that rationale and still dispute where the new lines should have been drawn, which is exactly what the legislative fight was about.

The fight took shape in the Senate Committee on Banking, Housing, and Urban Affairs under its chairman, Mike Crapo of Idaho, who introduced the bill as S. 2155 in late 2017. Crapo’s starting position reflected years of Republican criticism of Dodd-Frank, but the bill he produced was not the wholesale repeal that some in his party favored. It was a tailoring bill, and its contents reflected months of negotiation with a group of Democratic senators who had concluded that some relief for smaller institutions was justified. That negotiation is the hinge of the entire story, because in the Senate of the 115th Congress, with 51 Republicans and 49 members of the Democratic caucus, no banking bill could pass without Democratic votes. The chamber’s cloture rule, which required sixty votes to end debate on the legislation, meant that the majority needed at least nine votes from across the aisle for anything that could not be attached to a budget reconciliation measure, a procedural reality explained in the series guide to the filibuster and cloture. The tailoring bill was therefore designed from the start as a coalition product, and the coalition it assembled would become one of the most discussed features of the law.

What the fifty-billion-dollar line had meant in practice

To appreciate what section 401 changed, it helps to inventory what the old line actually triggered. A bank holding company that crossed 50 billion dollars in total consolidated assets entered a supervisory world designed for the largest and most complex firms in the country. It faced heightened capital and liquidity requirements beyond the generally applicable rules. It was subject to the Federal Reserve’s annual supervisory stress tests, with the results disclosed publicly. It had to conduct its own company-run stress tests under regulator-specified scenarios and maintain the governance machinery to oversee them. If publicly traded, it had to maintain a board-level risk committee with duties specified by regulation. It had to file resolution plans, the so-called living wills, describing how it could be resolved in bankruptcy without endangering the financial system. And it faced single-counterparty credit limits constraining its exposures to any one counterparty. Each of these requirements was defensible in isolation; together, they constituted a parallel regulatory regime entered by crossing a number.

The automaticity of the line was both its strength and its vulnerability. Its defenders argued that automaticity was the point: by triggering requirements as a matter of law, the line prevented supervisors from quietly relaxing standards during long calm periods when the political pressure to ease up would be strongest. The history of financial regulation offered ample precedent for that concern, since supervisory zeal has reliably faded as memories of the last crisis faded. An automatic threshold was a commitment device, a way for the Congress of 2010 to bind the regulators of 2018. Its critics replied that the commitment was miscalibrated: it bound regulators to treat a 55 billion dollar regional lender with a traditional deposit-and-loan model the same as a trillion-dollar global firm with a vast trading operation, and the compliance costs of that equivalence were not abstract. Building a stress-testing apparatus, a resolution-planning function, and a Volcker compliance program required specialized staff and systems whose cost did not scale down neatly with balance-sheet size.

The institutions caught by the line were, for the most part, regional banks that had grown past it through ordinary expansion rather than through the exotic activities that had caused the crisis. Their business models centered on taking deposits and making loans in defined geographic markets. Their failures, had they occurred, would have been painful for their communities and costly for the deposit insurance fund, but few analysts argued that any one of them could transmit distress through the financial system the way a large interconnected dealer could. That mismatch, between the systemic-risk rationale of the enhanced regime and the actual risk profiles of many firms it covered, was the substantive core of the case for re-tiering. The complete guide to the Dodd-Frank Act describes the original thresholds and the authorities they triggered, and the contrast between that architecture and the three-tier system of 2018 is the clearest single illustration of what the later statute did.

There was also a temporal dimension to the old line’s meaning. The 50 billion dollar figure had been chosen in 2010, in the immediate shadow of a crisis in which the financial system had nearly collapsed. Thresholds chosen in fear tend to be set with wide margins of safety, and the Congress of 2010 had every reason to prefer over-inclusion to under-inclusion. By 2018, the system had experienced nearly a decade of post-crisis regulation, capital levels at large firms were substantially higher, and the question facing Congress was whether the margin of safety should be maintained at its crisis-era width or narrowed to reflect calmer conditions and better-capitalized firms. The 2018 statute answered that question by narrowing the margin, and the March 2023 failures would later force a re-examination of whether the narrowing had gone too far.

The Politics of Getting to Sixty

What made community banks the center of the bill’s politics?

Community banks became the bill’s political center because they combined a sympathetic economic story with favorable geography. Thousands of small institutions held little systemic risk individually yet faced fixed compliance costs written for much larger firms, and their senators heard about it constantly. That made relief for the smallest lenders the least controversial part of the bill.

That relief became the foundation on which the broader coalition was built.

The coalition’s Democratic members were, with one independent exception, senators from states where community banking was a visible part of the economy and the political culture. Jon Tester of Montana, Heidi Heitkamp of North Dakota, Joe Donnelly of Indiana, Mark Warner of Virginia, Claire McCaskill of Missouri, and their colleagues represented the wing of the Democratic caucus most exposed to the argument that Dodd-Frank’s burdens were falling on the wrong institutions. Several faced difficult re-election campaigns in 2018 in states that Trump had carried, which their critics were quick to note, but the senators themselves framed their support as a matter of substance: the law as written was crushing small lenders, small lenders served communities that larger banks did not, and a tailored regime was better policy than a uniform one. Angus King of Maine, an independent who caucused with the Democrats, joined them, bringing the Democratic caucus supporters to seventeen members in total, sixteen Democrats plus King.

The opposition within the Democratic Party was led by Elizabeth Warren of Massachusetts, who made opposition to the bill a defining cause. Warren’s argument was that the bill’s threshold changes were not confined to genuinely small institutions, that the discretionary band between 100 and 250 billion dollars would leave large regional banks under-supervised, and that the political framing around community banks was cover for relief that would flow mainly to much larger firms. She was joined in opposition by the party’s leadership, including Chuck Schumer of New York and Sherrod Brown of Ohio, the ranking member of the Banking Committee, and by the progressive wing of the House. Outside Congress, the opposition was organized by groups including Americans for Financial Reform, whose executive director Lisa Donner was a prominent critic, and Better Markets. The argument among Democrats was therefore not about whether community banks deserved relief; even most opponents conceded that point. It was about how far up the size spectrum the relief should extend, and whether the supervisory discretion the bill granted the Federal Reserve could be trusted to hold the line for the institutions in the middle.

Republican support was broad but not quite unanimous, and the party’s leadership treated the bill as a major legislative priority. President Trump, who had campaigned on dismantling Dodd-Frank, endorsed the bill and signed it in a ceremony that presented it as the fulfillment of that promise, a framing that overstated what the text did but accurately captured the political symbolism. House Speaker Paul Ryan and Majority Leader Kevin McCarthy backed it, as did Jeb Hensarling of Texas, the chairman of the House Financial Services Committee, who had previously favored a far more aggressive rollback and accepted the Senate bill as the achievable version. On the regulatory side, Comptroller of the Currency Joseph Otting and Mick Mulvaney, who led the Consumer Financial Protection Bureau in an acting capacity during the bill’s passage, supported the tailoring approach. The politics, in short, arrayed a bipartisan coalition of the bill’s drafters against a largely Democratic opposition that warned the tailoring was miscalibrated, and both sides would later claim vindication in the events of 2023.

The Legislative Path: From Introduction to Signature

Senator Mike Crapo introduced S. 2155 in November 2017, and the bill’s journey from introduction to enactment illustrated how financial legislation moved through a Congress where the two parties disagreed sharply about the Dodd-Frank Act but agreed, in a bipartisan center, that its treatment of smaller institutions needed revision. The Senate Committee on Banking, Housing, and Urban Affairs took up the bill in December 2017, and the markup revealed the coalition’s shape before the floor fight did. Committee Democrats who would later vote for the bill on the floor supported it in committee, while the ranking member, Sherrod Brown of Ohio, opposed it, previewing the intraparty argument that would dominate the floor debate. The committee’s approval sent to the floor a bill that was already a negotiated product, not a partisan wish list, which distinguished it from the more aggressive House efforts to rewrite Dodd-Frank that had preceded it.

The contrast with the House’s earlier approach is instructive. Jeb Hensarling, chairman of the House Financial Services Committee, had spent years promoting a far-reaching alternative to Dodd-Frank, and the House had passed legislation reflecting that ambition. But the House bills could not clear the Senate, where the sixty-vote cloture requirement gave the Democratic minority a veto over anything that lacked bipartisan support. Crapo understood that arithmetic, and S. 2155 was written to satisfy it: a bill that gave Republicans the tailoring they wanted and gave a bloc of Democrats the community bank relief they could defend, while leaving the core Dodd-Frank authorities untouched so that the Democratic negotiators could answer the charge of repeal. The negotiation over the bill’s contents, conducted across party lines in late 2017 and early 2018, was the real legislative work; the floor votes ratified what the negotiation had produced.

Floor consideration in the Senate began in March 2018, and the debate organized itself around the two questions that had structured the negotiation. The bill’s supporters, led by Crapo and joined by the Democratic cosponsors, argued that the legislation was a targeted correction to a law whose burdens had been misallocated, and they returned relentlessly to the community bank story: small lenders in their states, crushed by compliance costs, consolidating or disappearing, taking credit access with them. The bill’s opponents, led by Elizabeth Warren, argued that the community bank provisions were the palatable wrapper around threshold changes that would benefit much larger institutions, and they returned relentlessly to the discretionary band: banks with hundreds of billions in assets, freed from automatic enhanced supervision, supervised only if the Federal Reserve chose to act. Both presentations were factually grounded in the bill’s text, which is why the debate was substantive rather than theatrical. The text did relieve community banks, and it did create the discretionary band. The disagreement was about which of those facts mattered more.

Amendments were offered and defeated along predictable lines, with the majority protecting the negotiated compromise against changes from either direction. Proposals to strengthen the bill’s consumer provisions or narrow its threshold changes failed, as did proposals to broaden its relief further. The management of the floor, keeping the coalition together through the amendment process, was the legislative achievement that made the 67 to 31 final vote possible. A bill that had been pulled apart by amendments from either flank would have lost votes it needed; a bill held together as a package could be presented, accurately, as the only bipartisan vehicle available. That framing, that this bill or no bill, was the Crapo strategy’s final element, and it worked.

The House’s handling of the legislation reflected the same arithmetic in reverse. Having passed its own more aggressive Dodd-Frank revisions, the House faced a choice: insist on its version and lose the Senate, or take up the Senate bill without amendment and send it to the president. The House chose the latter, passing S. 2155 unamended on May 22, 2018. The decision disappointed members who had wanted a more thorough rewrite, including some conservatives who considered the Senate bill inadequate, but the leadership, Speaker Paul Ryan and Majority Leader Kevin McCarthy, treated enactment of the achievable bill as preferable to the failure of the ambitious one. The absence of a conference committee, the usual mechanism for reconciling House and Senate versions, was itself a signal: the Senate bill was the final word, and the House’s role was ratification. President Trump’s signature on May 24, 2018, two days later, completed the process in just over six months from introduction, a pace that reflected both the years of groundwork that had preceded the bill and the determination of its supporters to finish before the legislative window narrowed.

The Roll Calls

The Senate passed S. 2155 on March 14, 2018, by a vote of 67 to 31, recorded as Roll Call Vote 54. The breakdown is worth setting out in full, because the composition of the majority is the single most cited fact about the bill’s politics and one of the most frequently misstated. Voting in favor were all 51 Republicans present, plus sixteen Democrats and one independent. The sixteen Democrats were Michael Bennet of Colorado, Tom Carper of Delaware, Chris Coons of Delaware, Joe Donnelly of Indiana, Maggie Hassan of New Hampshire, Heidi Heitkamp of North Dakota, Doug Jones of Alabama, Tim Kaine of Virginia, Joe Manchin of West Virginia, Claire McCaskill of Missouri, Bill Nelson of Florida, Gary Peters of Michigan, Jeanne Shaheen of New Hampshire, Debbie Stabenow of Michigan, Jon Tester of Montana, and Mark Warner of Virginia. The independent was Angus King of Maine. Voting against were 31 Democrats and one independent, Bernie Sanders of Vermont. Two senators did not vote: John McCain of Arizona and Martin Heinrich of New Mexico. The correct description of the Democratic support is therefore sixteen Democrats plus Senator King, or seventeen members of the Democratic caucus; the common misstatement that counts King as a seventeenth Democrat is wrong, and it has appeared often enough in commentary to merit the correction.

The size of the majority mattered procedurally as well as politically. Sixty-seven votes comfortably exceeded the sixty needed for cloture, which meant the bill’s supporters never faced the procedural cliff that kills most major legislation in the Senate. It also meant the bill had a veto-proof margin in the unlikely event of a presidential veto, though no veto was ever in prospect. The vote was the product of the negotiation described above: a bill written to attract a specific bloc of Democratic senators, passed with that bloc plus the nearly unified Republican conference. For a financial regulation bill in the polarized Congress of the late 2010s, the bipartisanship was genuine, and it was the reason the bill became law while more ambitious Republican repeal efforts did not.

The House took up the Senate bill without amendment and passed it on May 22, 2018, by a vote of 258 to 159. Every Republican present voted in favor except Walter B. Jones Jr. of North Carolina, and 33 Democrats joined them. The House vote drew less attention than the Senate’s, partly because the bill’s contents were fixed by then and partly because House passage was the expected final step, but the Democratic minority in favor is worth noting: it included members who would later defend the vote as support for community bank relief and members who would later face primary challenges over it. President Trump signed the bill two days later, on May 24, 2018, completing a legislative process that had run from Crapo’s introduction of S. 2155 in November 2017 through committee markup, floor debate, and final passage in just over six months, a brisk pace for a bill of its complexity.

With the politics established, the substance can be taken up in the order the statute presents it, beginning with the change that defined the law in public debate: the threshold for enhanced prudential standards.

The Democratic yes votes: a closer look at the coalition

The 16 Democrats and one independent who voted yes did not form an ideological bloc, and their individual situations explain more about the vote than any single theory. Several were among the most electorally vulnerable senators of the cycle. Heidi Heitkamp of North Dakota, Joe Donnelly of Indiana, Claire McCaskill of Missouri, and Bill Nelson of Florida all faced brutally competitive reelection contests in states that Donald Trump had carried, and each had reason to want the banking industry neutralized as a source of opposition spending. Jon Tester of Montana, a farmer who had built his political identity around rural economic issues, fit the same pattern: his state’s community banks were visible civic institutions, and their testimony about regulatory burden resonated with his constituents in ways that abstract defenses of Dodd-Frank did not. Doug Jones of Alabama, newly elected in a special election and facing an electorate far more conservative than his own voting record, had similar incentives to demonstrate independence from his party’s left flank.

Others came to the bill through longer-standing policy commitments. Mark Warner of Virginia, a former businessman with deep ties to the financial sector, had argued for years that Dodd-Frank’s thresholds were miscalibrated, and his vote reflected a substantive judgment rather than a tactical one. Tim Kaine of Virginia, Warner’s colleague, shared much of that outlook. Tom Carper and Chris Coons of Delaware represented a state whose economy was intertwined with banking, including the credit-card operations that federal law had drawn to Delaware decades earlier; their votes reflected a constituency in which financial regulation was not an abstraction but the local industry. Michael Bennet of Colorado, Gary Peters and Debbie Stabenow of Michigan, Maggie Hassan and Jeanne Shaheen of New Hampshire, and Joe Manchin of West Virginia each combined elements of both motives: genuine sympathy for the community-bank case and a political interest in the coalition the bill assembled.

Senator Angus King’s position was distinctive. As an independent who caucused with the Democrats, King was free from the party discipline that shaped his colleagues’ calculations, and his vote suggested that the bill’s merits could attract support outside the partisan frame entirely. His presence in the yes column also mattered arithmetically: with King, the Democratic caucus supplied 17 of the 67 yes votes, a quarter of the total, which allowed the bill’s supporters to describe it as genuinely bipartisan rather than as a Republican bill with a handful of Democratic defectors. The distinction mattered for the bill’s legitimacy and for the political cover it gave its supporters.

The no votes within the Democratic caucus were equally revealing. The 31 Democratic nay votes plus Senator Sanders included the party’s leadership, its most prominent progressives, and senators from states where the financial industry’s political weight was either hostile or irrelevant. Senator Schumer’s no vote, as Democratic leader, signaled that the party establishment opposed the bill even as a sixth of its caucus defected. The split mapped roughly onto the party’s ideological and geographic divides: progressives and members from deep-blue states opposed, moderates and members from red or purple states with significant community-banking sectors supported. That mapping would recur in later financial-regulation fights, making the 2018 roll call a useful predictor of where the party’s fault lines lay.

The Threshold: Section 401 and the Discretionary Band

Section 401 was the heart of the re-tiering. Before 2018, section 165 of the Dodd-Frank Act applied enhanced prudential standards automatically to every bank holding company with 50 billion dollars or more in total consolidated assets. The standards included supervisory stress tests, company-run stress tests, risk committee requirements, liquidity requirements, and resolution planning, and they applied by operation of law: cross the asset line, and the full apparatus engaged. After the 2018 statute, the automatic application began at 250 billion dollars in total consolidated assets, five times the old line. For bank holding companies between 100 and 250 billion dollars, the Federal Reserve received discretion to apply any or all of the enhanced standards on an institution-by-institution basis, but only upon a determination that doing so would promote financial stability or the safety and soundness of the institution. For bank holding companies below 100 billion dollars, the enhanced standards ceased to apply entirely, effective immediately upon enactment.

The timing provisions added a further layer. Institutions below the 100 billion dollar line were freed from the enhanced standards as soon as the president signed the bill. The remaining amendments, including the new 250 billion dollar automatic threshold and the discretionary band, took effect eighteen months after enactment, on November 24, 2019, giving the Federal Reserve time to write the implementing rules that would translate the statute’s framework into supervisory practice. The delayed effective date was a concession to the regulators, who needed to redesign a tiering system that had been built around a single line, and it became significant later, because it meant the full re-tiered regime had been in place for barely three years when the failures of March 2023 tested it.

Why did Congress give the Federal Reserve discretion instead of drawing one clean line?

Congress created the discretionary band because no single threshold commanded a majority. A line at 250 billion would have freed every bank entirely, which the Democratic negotiators rejected; a line at 100 billion would have preserved the regime for institutions the bill was meant to relieve. Discretion split the difference, converting an automatic rule into judgment for the middle tier.

The discretion, however, was not unbounded. The statute directed the Federal Reserve to consider factors including the institution’s capital structure, riskiness, complexity, financial activities, size, and any other risk-related factors the Board deemed appropriate, and to apply the standards only where doing so would promote financial stability or safety and soundness. In practice, this meant the Board had to build a new framework for deciding which institutions in the 100 to 250 billion dollar range would face which requirements, a framework it produced through rulemakings in 2019 that sorted large banking organizations into categories based on size, cross-jurisdictional activity, off-balance-sheet exposure, and reliance on short-term wholesale funding. The framework was complex, and its complexity would later become one of the criticisms leveled at the tailoring regime: that by replacing a simple automatic rule with a multi-factor discretionary system, Congress and the Board had made supervision harder to administer and easier to soften.

Two related provisions rode along with section 401 and are best understood as part of the same re-tiering. The threshold for mandatory company-run stress tests rose from 10 billion dollars to 250 billion dollars in total consolidated assets, a change that eliminated the requirement for a large group of mid-sized institutions and concentrated the testing regime on the largest firms. And the asset threshold for mandatory risk committees at publicly traded bank holding companies rose from 10 billion dollars to 50 billion dollars, relieving smaller public companies of a governance requirement that the bill’s drafters considered disproportionate to their risk profiles. Both changes followed the same logic as the main threshold: intensity of supervision should track the institution’s potential to threaten the system, and fixed low thresholds were a poor proxy for that potential.

The supporters’ case for the new threshold rested on a straightforward claim about proportionality. A regional bank with 60 or 80 billion dollars in assets, funded mostly by deposits and invested mostly in loans, did not pose the kind of risk that the enhanced prudential standards were designed to contain, and subjecting it to the full apparatus was pure deadweight cost. The opponents’ case rested on an equally straightforward claim about the value of automaticity. A bright line at 50 billion dollars might have been crude, but it was enforceable, predictable, and immune to the supervisory forbearance that had characterized the pre-crisis years; replacing it with discretion meant replacing a rule with a judgment, and judgments could be wrong. Both claims were plausible in 2018. The events of 2023 would give each side new evidence, and neither side would concede the other’s interpretation of it.

Proportionality Against Automaticity: The Intellectual Core of the Fight

Beneath the vote counts and the threshold numbers, the fight over the 2018 statute was a fight between two regulatory philosophies, and naming them clarifies everything else. The first philosophy, proportionality, holds that the intensity of regulation should vary continuously with the risk the regulated institution poses, and that any fixed threshold is at best a rough approximation of that variation. On this view, the pre-2018 regime’s 50 billion dollar line was indefensible in principle: it treated institutions of vastly different risk profiles identically, imposed identical compliance costs on firms with wildly different capacities to bear them, and generated no supervisory information commensurate with its cost for the smaller firms it captured. The remedy was tailoring, the deliberate matching of requirement to risk, through higher thresholds, discretionary bands, elective alternatives, and activity-based screens. The 2018 statute was proportionality enacted into law, and its defenders measured it by the precision of the match: did each provision’s relief reach the institutions whose risk did not justify the burden?

The second philosophy, automaticity, holds that the value of a regulatory rule lies substantially in its being a rule, in its applying without the exercise of discretion that can be captured, softened, or simply neglected. On this view, the pre-2018 regime’s bluntness was not a defect but a design feature: a bright line at 50 billion dollars was enforceable by outsiders, predictable for insiders, and immune to the supervisory forbearance that had allowed risks to accumulate before the 2008 crisis. Automaticity’s defenders did not deny that the line was crude; they denied that crudity was the relevant metric. The relevant metric was whether the system constrained risk-taking when it mattered, and automatic rules constrained it more reliably than discretionary ones because they did not depend on the assertiveness of particular supervisors in particular years. The 2018 statute, on this view, traded a reliable crude instrument for an unreliable precise one, and the trade was a mistake even if the precision was real.

The debate between these philosophies could not be resolved by the statute’s text, because the text embodied a compromise between them rather than a victory for either. The 250 billion dollar automatic threshold preserved automaticity for the largest firms; the discretionary band from 100 to 250 billion dollars embodied proportionality’s faith in supervisory judgment; the exemptions below 100 billion dollars reflected a judgment, shared by both sides, that automaticity’s benefits did not extend to the smallest institutions. The compromise’s vulnerability was the one the Barr review later identified: a system that depends on supervisory judgment depends on supervisors, and supervisors operate within institutional cultures that can drift. Proportionality’s defenders answered that any regulatory system depends on supervisors, including automatic ones, because even automatic rules require enforcement, interpretation, and updating. Automaticity’s defenders answered that dependence comes in degrees, and the 2018 law increased the degree past the prudent point.

The March 2023 failures did not settle the philosophical dispute, but they changed its terms by supplying a concrete case. Before 2023, the argument for automaticity was largely theoretical, a warning about what discretionary regimes might permit. After 2023, it had an exhibit: two banks in the discretionary band, supervised under the tailored framework, whose failures the Federal Reserve’s own review attributed in part to the tailoring approach. Before 2023, the argument for proportionality was largely an efficiency claim, a complaint about wasted compliance costs. After 2023, it had to answer the exhibit, and its answer, that the supervisory tools were adequate and unused, shifted the debate from the design of the framework to the performance of the supervisors. The philosophical dispute thus became an empirical one, and the empirical record, as the contested causal analysis showed, was genuinely ambiguous. That ambiguity is why the dispute persists, and why it will likely persist through the next re-tiering episode, whenever it comes.

Volcker: Section 203 and the Banking-Entity Exclusion

The Volcker Rule, section 619 of the Dodd-Frank Act, was among the most symbolically charged provisions of the post-crisis framework. Named for former Federal Reserve Chairman Paul Volcker, it prohibited banking entities from engaging in proprietary trading and from sponsoring or investing in hedge funds and private equity funds, the so-called covered funds. The rule’s purpose was to keep federally backed deposit-taking institutions out of speculative trading, and its implementation had been famously tortuous, generating thousands of pages of regulation and years of industry complaints about its complexity. For community banks, the Volcker Rule was a particular irritant, not because small banks were significant proprietary traders, most were not, but because the compliance apparatus required to demonstrate that they were not trading was itself expensive. A bank with no trading desk still had to maintain the policies, procedures, and documentation to prove the negative.

Section 203 of the 2018 statute addressed this complaint with a mechanism that is frequently misdescribed and worth stating precisely. The section did not merely exempt qualifying banks from the proprietary trading prohibition while leaving the rest of the rule in place. It excluded them from the definition of “banking entity” under section 13 of the Bank Holding Company Act, which is the definitional hook on which the entire Volcker Rule hangs. An institution that is not a banking entity is not subject to the proprietary trading ban and is not subject to the covered fund restrictions either; the whole rule falls away, not just one of its parts. That distinction matters because summaries of the law sometimes describe the change as a narrow trading exemption, when its legal effect was broader: for qualifying institutions, Volcker ceased to exist as a regulatory matter.

The qualification criteria were conjunctive, meaning both had to be satisfied. The institution’s total consolidated assets could not exceed 10 billion dollars, and its trading assets and liabilities could not exceed 5 percent of its total consolidated assets. The two-part test was designed to ensure that the exclusion reached only institutions that were both small and genuinely uninvolved in trading: a small bank with an outsized trading book would fail the second prong, while a large bank with minimal trading would fail the first. The provision took effect immediately upon enactment, and it applied to banks and their holding companies alike, provided both tests were met.

The policy argument for the exclusion was the tailoring rationale in its purest form. If an institution held less than 10 billion dollars in assets and devoted less than 5 percent of its balance sheet to trading, the Volcker Rule’s prohibitions were guarding against a risk the institution did not pose, while its compliance costs were real and regressive, falling hardest on the smallest firms. The counterargument, pressed by the bill’s opponents, was that the exclusion’s breadth, the whole rule rather than just the trading ban, was disproportionate to the problem, and that the covered fund provisions in particular had value even for small institutions as a guard against the migration of risky activities into fund structures. As with the threshold change, the disagreement was not about whether small banks were systemically dangerous; everyone agreed they were not. It was about whether the compliance relief was worth the loss of the rule’s prophylactic coverage, and about whether a 10 billion dollar line was the right place to draw it.

What section 203 did not do also matters, because the negative space defines the re-tiering as clearly as the positive. Banks above the 10 billion dollar line remained fully subject to the Volcker Rule, both the proprietary trading prohibition and the covered fund restrictions. The rule’s implementing regulations, the product of years of interagency work, continued to apply to every institution that did not meet the exclusion criteria. And the largest trading houses, the institutions whose activities had motivated the rule in the first place, were untouched. The Volcker change was therefore a paradigmatic coverage change: the substance of the rule, what it prohibited and why, was left exactly as Congress had written it in 2010, while the set of institutions required to live under it shrank.

The Covered Fund Question Inside the Volcker Exclusion

The Volcker Rule’s covered fund provisions, the restrictions on sponsoring or investing in hedge funds and private equity funds, received less public attention than the proprietary trading ban but were the more consequential half of the rule for the financial system’s structure. The trading ban addressed what banks did with their own balance sheets; the covered fund provisions addressed the channels through which risk could migrate out of the banking system into less regulated vehicles while remaining connected to banks through sponsorship, investment, or implicit support. When section 203 excluded qualifying small banks from the banking entity definition, it removed both halves of the rule, and the covered fund half of that removal deserves separate scrutiny, because the policy arguments for and against it differ from those surrounding proprietary trading.

The argument for removing the covered fund restrictions from small banks rested on the same proportionality logic as the rest of the provision. A bank with less than 10 billion dollars in assets and minimal trading activity was not a meaningful sponsor of hedge funds or investor in private equity, and the compliance apparatus required to monitor covered fund relationships, to distinguish permitted investments from prohibited ones, to document every fund interaction, was disproportionate to the risk. The covered fund rules, like the trading ban, had been written with the large trading houses in mind, and their application to small banks produced paperwork without protection. On this view, the exclusion simply recognized that the provision’s purposes were not served by its application to institutions that did not engage in the covered activities.

The counterargument focused on the migration channel rather than the institution’s own activities. The covered fund restrictions served a systemic purpose beyond constraining any single bank: they limited the banking system’s entanglement with the less regulated fund sector, reducing the channels through which stress in one part of the system could propagate to federally backed institutions. A small bank’s individual fund investments might be immaterial, but the principle that banking entities should not sponsor or invest in covered funds was, on this view, a structural safeguard whose value did not scale down with the institution’s size. Removing it for a category of institutions, even a low-risk category, created an exception to a structural rule, and structural rules lose their force through exceptions. The critics did not claim that small banks’ fund activities posed imminent danger; they claimed that the architecture of separation was worth preserving intact, and that the compliance costs, while real, were the price of the architecture.

The provision’s defenders had a reply, and it returned to the qualifying criteria. The two-part test, size plus trading activity, ensured that the exclusion reached only institutions whose fund activities, if any, were de minimis, and the Federal Reserve retained authority over the safety and soundness of every excluded institution through the ordinary supervisory process. An excluded bank that developed a significant fund business would trip the trading activity test or draw supervisory attention long before its activities became systemically relevant. The architecture of separation, on this view, was preserved where it mattered, at the institutions whose fund activities could actually transmit stress, and relaxed where it did not matter, at institutions whose activities could not. Whether that confidence in the screens and the supervision was justified was, like so much about the 2018 law, a question the quiet years did not test and the 2023 failures tested only indirectly, since neither failed bank’s story centered on covered fund activities.

The Community Bank Leverage Ratio: Section 201

If section 401 was the most debated change and section 203 the most symbolically charged, section 201 was the most technically elegant, and it embodied the tailoring philosophy in a different way. Rather than exempting small banks from capital rules, it offered them a simpler way to satisfy the underlying safety objective. The provision directed the federal banking agencies to establish a Community Bank Leverage Ratio, a single leverage measure that a qualifying community bank could elect to use in place of the full risk-based capital framework. A bank that maintained the ratio would be deemed to satisfy the generally applicable capital requirements and the well-capitalized standards of the prompt corrective action regime, without computing risk-weighted assets at all.

The statute set the parameters within which the agencies were to work. It directed them to set the ratio at a level between 8 and 10 percent, and the final rule, issued on October 29, 2019, set it at 9 percent of Tier 1 capital to average total consolidated assets. The qualifying criteria tracked the Volcker exclusion’s approach of combining a size test with an activity test: total consolidated assets below 10 billion dollars, off-balance-sheet exposures no greater than 25 percent of average total consolidated assets, trading assets and liabilities no greater than 5 percent of total consolidated assets, and exclusion of advanced-approaches banking organizations, the largest firms subject to the most complex capital rules. A bank whose leverage ratio slipped was given a two-quarter grace period to restore compliance, provided the ratio remained at or above 8 percent, a cushion designed to prevent the election from becoming a trap during temporary balance sheet fluctuations.

The appeal of the leverage ratio option lay in what it eliminated. The risk-based capital framework requires banks to assign risk weights to every category of asset, maintain the systems to compute those weights, and report the results in exhaustive regulatory filings. For a community bank whose portfolio consisted of straightforward loans, the exercise produced little supervisory information at considerable cost. The leverage ratio replaced all of that with a single question: did the bank hold enough Tier 1 capital against its total assets? The 9 percent level was set deliberately above the generally applicable leverage requirement, so the simplicity came at the price of holding more capital, a trade that the provision’s designers considered appropriate. Banks that preferred the risk-based framework, perhaps because their asset mix made it advantageous, remained free to use it; the leverage ratio was an election, not a mandate.

Critics of the provision raised two concerns. The first was that a simple leverage ratio was blind to asset risk, and that a bank could satisfy the 9 percent threshold while holding a portfolio of unusually risky loans, a blindness the risk-based framework was designed to correct. The provision’s defenders answered that the qualifying criteria, particularly the limits on off-balance-sheet exposure and trading activity, screened out the institutions most likely to game a simple ratio, and that examiners retained full authority to criticize risky lending through the ordinary supervisory process. The second concern was about the calibration of the 9 percent level itself, with some analysts arguing it was set high enough to discourage uptake and others arguing it was the minimum consistent with safety. The provision’s history after enactment, with community banks adopting the election in significant numbers, suggested the calibration was at least workable, though the debate over the right level continued in the rulemaking comments and the trade press.

Mortgages and Reporting: Sections 101 and 104

The mortgage provisions of the 2018 statute addressed one of the most persistent complaints of community lenders: that the post-crisis mortgage rules, however well designed for the large originators whose practices had fueled the crisis, imposed disproportionate costs on small institutions that held the loans they made. Two sections carried this part of the re-tiering, one dealing with the qualified mortgage standard and one with Home Mortgage Disclosure Act reporting, and both followed the same pattern of carving portfolio lenders out of obligations written with a different business model in mind.

Section 101 created a qualified mortgage safe harbor for certain residential mortgage loans originated and retained by small depository institutions and credit unions. Under the Dodd-Frank Act’s ability-to-repay framework, a lender that originated a qualified mortgage received either a safe harbor or a rebuttable presumption of compliance with the requirement to verify the borrower’s ability to repay, a valuable legal protection in a litigious market. The 2018 provision extended that protection to loans originated by institutions with less than 10 billion dollars in total consolidated assets, provided the institution retained the loan in portfolio rather than selling it into the secondary market. The qualifying loans had to meet familiar guardrails: no negative amortization features, no interest-only terms, limits on points and fees, and a documented process in which the lender considered and verified the borrower’s debt, income, and financial resources. The safe harbor generally survived only while the originating institution held the loan, with transfer permitted only in narrow circumstances, a condition that tied the legal protection to the economic alignment the provision was meant to reward.

The logic of the portfolio condition was the provision’s most defensible feature and the one its supporters emphasized. A lender that kept the loan on its own books had every incentive to underwrite carefully, because it would bear the loss if the borrower defaulted; the originate-to-distribute model, in which the lender’s profit came from volume rather than loan performance, was the model that had produced the crisis-era abuses. By conditioning the safe harbor on retention, the statute aligned the legal benefit with the economic incentive, extending relief precisely to the lenders whose business model already embodied the prudence the ability-to-repay rule was meant to enforce. The provision’s critics did not dispute the logic so much as its boundaries, questioning whether the 10 billion dollar line was the right cutoff and whether the guardrails were sufficient for the minority of portfolio lenders whose underwriting was less careful than the idealized community bank of the supporters’ rhetoric.

Section 104 addressed a different mortgage-related burden: the expanded data reporting required by the Home Mortgage Disclosure Act. The Consumer Financial Protection Bureau’s 2015 HMDA rule had added a large number of new data fields to the reporting obligations of covered lenders, and the compliance systems required to capture and report the expanded data were expensive to build and maintain. Section 104(a) provided partial relief: insured depository institutions and credit unions that originated fewer than 500 closed-end mortgages or 500 open-end lines of credit in each of the two preceding calendar years were exempted from reporting the expanded data fields added by the 2015 rule. The exemption was partial in a precise sense: the 22 data points that lenders had reported before the 2015 expansion continued to be required, so the public record of who was lending to whom did not go dark. It was limited to depository institutions and credit unions, excluding nonbank lenders entirely, and it excluded institutions with poor Community Reinvestment Act ratings, a condition that tied the relief to a record of serving the institution’s community.

Two common misdescriptions of this provision should be corrected, because both have circulated widely. The first is the claim that the provision raised the closed-end loan reporting threshold from 25 to 500 originations for a period of two years. That description is wrong on both the mechanics and the duration: the provision exempted qualifying institutions from the expanded fields added in 2015 while leaving the pre-existing reporting intact, and it contained no two-year sunset. The second misdescription is the suggestion that the provision gutted mortgage data collection. The 22 pre-existing data points, which included the core information about loan amounts, borrower demographics, and lending patterns that researchers and regulators had relied on for decades, remained fully reportable. What the provision reduced was the marginal burden of the newest and most granular fields, for the smallest lenders, under conditions designed to preserve the data’s fair-lending value. Whether that balance was correctly struck was debated at the time and has been debated since, but the debate should proceed from what the provision did, not from the caricatures.

The Targeted Provisions: Custody, Exams, and Municipal Securities

Beyond the five central changes, the 2018 statute contained a series of targeted provisions for particular institution types, each reflecting a specific industry complaint that had found a legislative champion. Three of them illustrate the re-tiering logic in miniature, and each is worth understanding on its own terms.

Section 402 addressed the supplementary leverage ratio as applied to custody banks, the institutions whose business consists primarily of safekeeping assets for institutional clients. Custody banks hold large volumes of deposits at central banks, not as a lending resource but as a mechanical feature of their business model: client cash awaiting investment sits as deposits, which the custody bank places at the central bank. Under the pre-2018 supplementary leverage ratio, those central bank deposits counted as leverage exposure, which meant the custody banks’ business model generated a capital requirement against assets that were, by common consent, among the safest on any balance sheet. Section 402 excluded central bank deposits from the supplementary leverage ratio for qualifying custody banks, aligning the capital treatment with the economic reality. The Federal Reserve’s final rule implementing the provision was issued on November 19, 2019, and took effect on April 1, 2020. The provision was narrow, technical, and uncontroversial among its supporters and most of its critics; its significance for this article is as a clean example of tailoring, a rule adjusted to fit the risk profile of a specific business model rather than repealed or left miscalibrated.

Section 210 raised the asset threshold for the extended examination cycle from 1 billion dollars to 3 billion dollars. Under the pre-2018 framework, qualifying well-capitalized and well-managed institutions below the threshold could be examined on an eighteen-month cycle rather than the standard twelve-month cycle, a scheduling accommodation that reduced the examination burden on small institutions without, in the regulators’ judgment, reducing supervisory effectiveness. The 2018 statute extended that accommodation to institutions up to 3 billion dollars, and the agencies’ final rule, issued on December 18, 2018 and effective January 28, 2019, implemented the change. The provision is worth noting for a further reason: Congress later revisited the line again, and a subsequent statute, Public Law 119-101, enacted in July 2026, raised the threshold to 6 billion dollars. The 2018 change should therefore be understood as what the 2018 Congress set, not as the final word on the subject, and the later revision illustrates the amendment stage continuing its work.

Section 403 addressed the liquidity coverage ratio’s treatment of municipal securities. The post-crisis liquidity rules classified assets by their liquidity characteristics, with the most liquid assets qualifying as high-quality liquid assets that banks could count toward their liquidity buffers. Municipal obligations, despite their generally strong credit quality and active markets, had been treated less favorably than their characteristics warranted, in the view of the provision’s supporters, which raised borrowing costs for the state and local governments that issued them. Section 403 directed that qualifying municipal obligations, those that were investment grade, liquid, and readily marketable, be treated as Level 2B high-quality liquid assets. The change was a modest recalibration of the liquidity framework’s asset classifications, and like the custody provision, it illustrated the statute’s method: not the repeal of the liquidity rules, but the adjustment of their application to a specific asset class whose treatment the drafters considered miscalibrated.

The Rulemaking Years: Building the Tailored Framework

Enactment was the beginning of the implementation, not its end, and the years from 2018 to 2020 were consumed by the rulemakings that translated the statute’s framework into supervisory practice. The most consequential of these was the Federal Reserve’s construction of the tailoring framework for the discretionary band, the system by which the Board would decide which institutions between 100 and 250 billion dollars faced which enhanced standards. The statute had given the Board factors to consider, including capital structure, riskiness, complexity, financial activities, and size, and had authorized application of the standards only where doing so would promote financial stability or safety and soundness. Turning those instructions into rules required the Board to define categories of large banking organizations sorted by risk indicators beyond mere asset size: cross-jurisdictional activity, off-balance-sheet exposure, reliance on short-term wholesale funding, and nonbank assets, alongside total consolidated assets. The resulting framework, finalized in the fall of 2019, sorted firms into tiers with graduated requirements, so that the most complex and interconnected firms in the band faced standards closer to those applied above 250 billion dollars, while simpler domestic firms faced lighter ones.

The framework’s complexity was both its point and its vulnerability. Its point was precision: supervision calibrated to the actual risk profile of each institution rather than to a single asset number. Its vulnerability, identified later in the Barr review’s fourth takeaway, was that complexity made the system harder to administer, easier to misunderstand, and more susceptible to a supervisory culture that favored forbearance. The pre-2018 regime’s virtue had been its bluntness: cross the 50 billion dollar line and the full apparatus applied, no judgment required. The post-2018 regime replaced that bluntness with a multi-factor sorting system that demanded continuous supervisory judgment, and the quality of the supervision therefore depended on the quality and assertiveness of the supervisors in a way the old regime had not. That dependence would become central to the debate after 2023, when critics argued the framework had been designed to fail softly and defenders argued it had been designed to be used firmly.

The other rulemakings proceeded on their own tracks. The Community Bank Leverage Ratio required the banking agencies to set the ratio within the statute’s 8 to 10 percent range and define the qualifying criteria; the final rule of October 29, 2019, set the ratio at 9 percent and established the asset, exposure, trading, and advanced-approaches screens described earlier. The custody bank provision of section 402 required a Federal Reserve rule excluding central bank deposits from the supplementary leverage ratio; the final rule of November 19, 2019, effective April 1, 2020, implemented it. The extended examination cycle of section 210 was implemented by an interagency final rule of December 18, 2018, effective January 28, 2019, raising the threshold from 1 to 3 billion dollars. The mortgage provisions required action by the Consumer Financial Protection Bureau, which undertook the rulemakings and guidance necessary to implement the qualified mortgage safe harbor and the partial HMDA exemption. Each rulemaking drew comments, each made technical choices within the statute’s parameters, and each became part of the operating reality that the industry inhabited from 2020 forward.

The rulemaking period also revealed the statute’s dependence on agency good faith, a dependence its drafters had accepted and its critics had flagged. Where the statute granted discretion, as in the 100 to 250 billion dollar band, the agencies’ choices determined the law’s practical meaning. Where the statute set ranges, as in the 8 to 10 percent leverage ratio band, the agencies’ calibration determined the provision’s generosity. A different set of regulators could have implemented the same statute more strictly or more leniently, and the knowledge of that contingency hung over the framework from the start. The Barr review would later argue that the contingency had resolved in the lenient direction, a claim the review’s critics disputed but could not dismiss, because the structure of the discretion made the claim at least plausible.

The Discretionary Band as a Supervisory Experiment

The 100 to 250 billion dollar discretionary band was the 2018 statute’s most original institutional creation, and it is worth considering as an experiment in regulatory design, whatever one’s view of its results. Before 2018, American bank supervision above the 50 billion dollar line had been rule-driven: the statute specified the requirements, the asset line triggered them, and the supervisors implemented what the law commanded. The discretionary band introduced a different model, one in which the statute specified the considerations and the supervisors supplied the judgment, institution by institution. The experiment’s premise was that supervisors, equipped with detailed knowledge of each firm’s risk profile, could calibrate requirements more precisely than any fixed line, applying the full apparatus where the risk warranted it and withholding it where the risk did not. The experiment’s risk was that the judgment would be exercised weakly, inconsistently, or under political and industry pressure, and that the band would become in practice a zone of relief rather than a zone of calibration.

The Federal Reserve’s implementation of the band, through the tailoring framework finalized in the fall of 2019, was an attempt to discipline the discretion with structure. By sorting firms into categories based on size, cross-jurisdictional activity, off-balance-sheet exposure, wholesale funding reliance, and nonbank assets, the framework gave supervisors a rule-like scaffolding within which to exercise judgment, reducing the arbitrariness that pure discretion would have entailed. The categories determined which requirements applied presumptively, leaving the Board room to adjust at the margins. In design terms, the framework was a hybrid: more flexible than the old automatic line, more structured than unguided discretion. Its designers considered the hybrid the best of both philosophies, preserving proportionality’s precision while constraining its vulnerability to forbearance.

The Barr review’s assessment of the experiment was skeptical. The review found that the tailoring approach had impeded effective supervision by reducing standards, increasing complexity, and promoting a less assertive supervisory approach, a verdict that treated the hybrid’s complexity as a cost rather than a safeguard. Complexity, on this account, did not discipline discretion; it obscured it, making it harder for supervisors, managers, and outside observers to know what was required and whether it was being enforced. The less assertive supervisory approach was the mechanism by which the complexity translated into outcomes: faced with a multi-factor framework and a culture that favored restraint, supervisors chose the less demanding path. The review’s critics, including the statute’s defenders, disputed the attribution, arguing that the supervisory culture had independent causes and that the framework’s structure was sound. But the dispute itself revealed the experiment’s inherent difficulty: when a discretionary regime underperforms, its designers can always blame the exercise of the discretion rather than its grant, and its critics can always blame the grant rather than the exercise, and no outcome cleanly distinguishes the two.

The deeper lesson of the experiment concerns the relationship between statutory design and supervisory capacity. A discretionary regime demands more of supervisors than an automatic one: more analysis, more judgment, more willingness to act on the analysis against industry resistance. The 2018 Congress created such a regime while the supervisory agencies were simultaneously experiencing the cultural and leadership shifts that the Barr review documented, and the combination produced less supervision than the statute’s authors had contemplated. Whether a different supervisory culture would have made the band work as designed is the counterfactual at the heart of the contested causal debate, and it is untestable. What can be said is that the experiment demonstrated, at a minimum, that discretionary tailoring is only as strong as the supervision that implements it, a proposition both sides of the debate could endorse even as they drew opposite conclusions from it.

After Enactment: The Quiet Years

From the completion of the major rulemakings in 2020 through the end of 2022, the re-tiered regime operated without a systemic test, and the period’s quietude shaped how the statute was evaluated before the failures. Community banks adopted the Community Bank Leverage Ratio election in significant numbers, suggesting that the 9 percent calibration had found the workable middle its designers intended: simple enough to attract users, demanding enough to satisfy safety concerns. Institutions that qualified for the extended eighteen-month examination cycle used it. The qualified mortgage safe harbor for portfolio loans became part of small lenders’ compliance toolkit. The Volcker exclusion removed a documented irritant for banks below the 10 billion dollar line with minimal trading activity. By the measures its supporters had proposed, the law was working: compliance costs fell for the institutions it targeted, and no supervisory crisis emerged to suggest the relief had been reckless.

The quiet years also saw the consolidation trend continue, which complicated the supporters’ narrative without refuting it. Community bank numbers kept declining through mergers, driven by forces larger than any single statute: technology costs, demographic shifts, competition from nonbank lenders, and the long-run economics of scale in banking. Defenders of the 2018 law argued that the decline would have been steeper without the relief, a counterfactual that could not be tested directly. Critics argued that the law’s benefits flowed disproportionately to larger institutions while the smallest continued to disappear, a claim that shifted the evaluation from the law’s intent to its incidence. The data could sustain either emphasis, which is why the period produced no consensus about the law’s effects, only a stalemate of competing interpretations that the events of 2023 would break open.

Supervisory practice during the quiet years is harder to assess from the public record, but the Barr review’s retrospective account described a supervisory stance that had grown less assertive, and attributed part of that shift to the tailoring framework’s complexity and reduced standards. Whether the softening was caused by the 2018 statute, by the change in administration and regulatory leadership, by the long expansion’s complacency, or by some combination of all three, was the question the review’s fourth takeaway raised without fully resolving. What is clear is that the re-tiered regime’s first years coincided with a period in which supervisors identified problems at growing regional banks but pursued them slowly, through the graduated machinery of supervisory findings rather than enforcement, and that this pattern would look, in retrospect, like a missed opportunity to test whether the discretionary band’s authority would be used when it mattered.

What the Bureau Kept: Consumer Provisions and the Untouched CFPB

One of the most persistent myths about the 2018 statute is that it weakened the Consumer Financial Protection Bureau. The myth is understandable: the Bureau was the most politically contested creation of the Dodd-Frank Act, the Trump administration was openly hostile to it, and the 2018 law was signed by a president who had promised to dismantle the post-crisis framework. But the myth is false, and its falsity is another illustration of the re-tiering principle. The 2018 statute did not alter the Bureau’s structure, its funding, or its authorities in any respect. The director’s powers, the Bureau’s independent funding through the Federal Reserve, its rulemaking and enforcement jurisdiction: all of it survived the 2018 law exactly as Dodd-Frank had established it. The series account of the Bureau’s creation and powers describes the authorities that the 2018 Congress left alone, and the contrast between the political rhetoric surrounding the bill and the statutory text on this point is one of the sharpest in the entire episode.

What the statute did do on the consumer side was add new protections and assign the Bureau new implementation work. Section 309 imposed a seasoning requirement on certain refinancing loans guaranteed by the Department of Veterans Affairs, a response to a pattern of rapid, repeated refinancings that stripped equity from veteran borrowers through fees while providing little benefit. The provision required a minimum period between the original loan and the refinance and imposed other guardrails designed to ensure that VA refinancings served the borrower’s interest. Other provisions of the statute addressed the treatment of private student loan debt following the death or total and permanent disability of the borrower, requiring lenders to release cosigners and discharge obligations in circumstances where collection had been a source of sustained complaint. These were consumer-facing additions, not deregulatory subtractions, and they complicated the simple story that the 2018 law moved in only one direction.

The statute also gave the Bureau new work in implementing its own provisions. The HMDA partial exemption, the qualified mortgage safe harbor, and several of the consumer provisions required rulemakings or interpretive guidance, and the Bureau undertook that work in the years following enactment. The assignment is worth noting because it underscores the institutional reality that the 2018 Congress accepted: the Bureau it declined to restructure was the Bureau it trusted to implement the new law’s consumer provisions. Whatever the administration’s posture toward the agency, the statute treated it as a going concern with work to do.

How did supporters describe the bill without calling it a rollback?

Supporters rarely used the word rollback, and the avoidance was deliberate. The bill’s title spoke of economic growth, regulatory relief, and consumer protection, and its defenders framed every provision as tailoring: matching regulatory intensity to institutional risk. Crapo called it right-sized regulation; Tester and Heitkamp called it relief for Main Street lenders.

The administration called it the fulfillment of the promise to help community banks. The vocabulary mattered because it defined the terms of the later debate: if the law was tailoring, then the 2023 question was whether the tailoring had been calibrated correctly, a technocratic dispute. If it was a rollback, then the 2023 question was whether deregulation had caused the failures, a political one.

Why did the consumer title matter politically to the bill’s Democratic supporters?

The statute added VA refinance seasoning rules and student-loan discharge protections while leaving every CFPB authority intact, and those additions mattered politically because they gave Democratic supporters a consumer-protection record to cite against the charge that the bill served only the banking industry.

The political logic was straightforward. Senators like Heitkamp, Donnelly, Tester, and Manchin, most of them facing difficult reelection contests in states Donald Trump had carried, needed to show that their vote for regulatory relief did not mean abandoning consumers. The VA seasoning provision, which protected veterans from equity-stripping refinance churn, and the student-loan provisions, which addressed genuinely sympathetic cases of borrower death and disability, were difficult for opponents to attack on substance. Senator Warren’s opposition did not deny their value; her argument was that they were sweeteners attached to a bill whose core was dangerous. That framing, sweeteners versus core, structured the entire Democratic debate over the statute, and the 16 Democrats who voted yes effectively answered that the package, taken whole, was worth supporting.

Veterans, students, and homeowners: the consumer title in detail

The consumer-facing provisions of the 2018 statute merit a closer look, because they complicate every simple story about the bill and because they reveal how bipartisan coalitions are actually assembled. Section 309’s protection for veterans addressed a specific abuse that had drawn bipartisan concern. Lenders specializing in VA-guaranteed loans had been soliciting veterans for rapid serial refinancings, sometimes within months of the previous loan, each refinance generating new fees and stripping equity that the veteran had barely begun to accumulate. The practice, known as churning, was profitable for the lenders and costly for the borrowers, and it also imposed costs on the VA guaranty program, which bore the credit risk of the churned loans. Section 309 imposed seasoning requirements, mandating a minimum period before a VA loan could be refinanced, which broke the economics of the churn without restricting legitimate refinancings undertaken for genuine borrower benefit. The provision’s design reflected a sophisticated understanding of the abuse: it did not ban VA refinancing, which serves veterans well when rates fall, but required the loan to age before it could be refinanced, ensuring that each transaction had to deliver real value to justify its costs.

The student-loan provisions addressed a different set of sympathetic cases. Private student loans had developed a reputation for harsh treatment of borrowers in distress, partly because the loans were difficult to discharge and partly because servicers applied inconsistent standards. The statute added protections around the discharge of private student loans in cases of borrower death or disability, requiring clearer treatment of cosigners and improving transparency around the conditions for cosigner release. A cosigner, often a parent, who had guaranteed a child’s loan could previously find the loan’s terms unforgiving when tragedy struck; the new provisions required lenders to address those situations more humanely and to disclose release options more clearly. These were retail-level reforms, invisible in the macroeconomic debate over thresholds but tangible for the families they reached.

The political function of these provisions was as important as their substantive content. In the negotiations that produced the bill, the consumer title gave the Democratic supporters a record to run on: they could point to specific protections for veterans and students that they had voted to enact, countering the charge that their vote served only the banking industry. The provisions also gave the bill’s Republican sponsors a broader story to tell, one in which regulatory relief for lenders was paired with concrete protections for borrowers. Whether the pairing was principled or merely transactional depends on one’s view of legislative bargaining, but its effect on the vote count is not in dispute. Several of the Democratic yes votes cited the consumer provisions explicitly in explaining their positions, and the provisions’ presence in the bill was a necessary condition of the coalition, even if it was not a sufficient one.

What did opponents mean when they called the bill a Wall Street giveaway?

Opponents used the phrase to argue that threshold relief reached far beyond community banks to institutions holding hundreds of billions of dollars in assets. Supporters answered that the largest firms saw no change, since automatic coverage stayed at 250 billion dollars, and that the most generous provisions were confined to the smallest lenders.

The framing battle over the bill was fought with labels that each captured a fragment of the truth. “Wall Street giveaway” was the opposition’s most potent phrase, and its power came from the threshold arithmetic: a statute sold as community-bank relief also moved the automatic line for firms with 200 billion dollars in assets, and no ordinary use of language puts such firms in the same category as a rural community bank. The phrase’s weakness was its implication about the largest firms: the money-center banks that most Americans picture when they hear “Wall Street” experienced no relief whatsoever, since the enhanced regime continued to apply to them automatically and in full. A more precise version of the opposition’s claim would have been “large regional bank relief,” but precision is rarely the point of political language.

The supporters’ preferred label was “regulatory relief,” and it had the symmetric problem. Relief was real for the institutions it reached, but the phrase implied a burden lifted across the board, when in fact the statute’s benefits were carefully tiered and its largest beneficiaries in dollar terms were mid-sized firms rather than the smallest ones. The community banks that supplied the bill’s sympathetic imagery received the CBLR option, the Volcker exclusion, the mortgage provisions, and the exam-cycle extension; the regional banks that supplied much of the lobbying energy received the threshold re-tiering. Both groups got something, but they did not get the same thing, and the single label obscured the difference.

Underneath the labels lay a genuine disagreement about what the financial system needed. The bill’s supporters believed that the post-crisis regime had overshot, applying crisis-era stringency to institutions and activities that did not warrant it, and that the resulting compliance costs were distorting the structure of American banking by accelerating consolidation. The bill’s opponents believed that the post-crisis regime had approximately hit its target, that the compliance costs were the price of safety, and that any narrowing of the perimeter would be exploited, first by the institutions directly relieved and eventually by the political dynamics that relief would unleash. Those two views could not be reconciled by evidence available in 2018, because they were views about the future: about whether supervisors would use their restored discretion wisely, about whether the next crisis would come from the institutions the bill had relieved, about whether the political system could hold the new lines. The March 2023 failures supplied the first real test of those predictions, and the two sides read the results as differently as they had read the bill.

What Nothing Changed: The Authorities Left Intact

The re-tiering claim can be tested most directly by inventorying what the 2018 statute did not touch, because the list is long and it includes every authority that a genuine repeal would have had to reach. The Orderly Liquidation Authority, Title II of Dodd-Frank, which gives the Federal Deposit Insurance Corporation the power to resolve a failing systemically important financial company outside of bankruptcy, survived intact, and it remained unused as of the period this article covers. The Financial Stability Oversight Council’s authority to designate nonbank financial companies and financial market utilities for heightened supervision survived intact. The Bureau’s structure, funding, and authorities survived intact, as described above. The derivatives clearing mandate of Title VII, the requirement that standardized swaps be cleared through central counterparties and traded on transparent platforms, survived intact. The Volcker Rule continued to apply in full to every banking entity above the 10 billion dollar exclusion line, including all of the largest trading houses. And the resolution planning requirement of section 165(d), the living wills that the largest firms filed with the regulators, continued to apply to firms above the new 250 billion dollar threshold.

The significance of this inventory is not merely defensive, though it does answer the repeal question definitively. It also defines the character of the 2018 Congress’s intervention. A legislature that wanted to repeal Dodd-Frank would have attacked the authorities themselves: the liquidation power, the designation power, the Bureau, the clearing mandate. The 115th Congress attacked none of them. What it did instead was adjust the sorting mechanism, moving lines, creating a discretionary band, and carving out categories of institutions and activities. The result was a Dodd-Frank Act whose substantive commands were unchanged but whose reach was narrower and more graduated. That is the re-tiering, and it is the reason the repeal debate and the calibration debate must be kept distinct. One can believe the calibration was wrong, that the 250 billion dollar line was too high or the discretionary band too permissive, while fully accepting that no repeal occurred. Much of the public argument has failed to hold both thoughts at once.

The inventory also clarifies what the 2018 statute was not. It was not a response to a demonstrated failure of the Dodd-Frank framework; no major institution had failed in the years preceding its passage, and the framework’s defenders could plausibly claim it was working. It was not a crisis measure. It was a judgment, made in calm economic conditions, that the framework’s coverage was broader than its purposes required, and like all such judgments, it could only be evaluated in retrospect. The retrospective evaluation arrived sooner and more dramatically than anyone expected.

March 2023: The Failures That Tested the Framework

In March 2023, less than five years after the 2018 statute’s enactment, the re-tiering faced the test its critics had predicted and its supporters had discounted. Two large regional banks failed within days of each other, in the second and third largest bank failures in American history at the time, and both sat squarely in the zone the 2018 law had re-tiered: above the old 50 billion dollar automatic threshold, below the new 250 billion dollar one.

Silicon Valley Bank, headquartered in Santa Clara, California, was closed on March 10, 2023, by the California Department of Financial Protection and Innovation, which appointed the Federal Deposit Insurance Corporation as receiver. The bank held approximately 209 billion dollars in assets at year-end 2022, making it the sixteenth largest bank in the country by assets and, at the time of its closure, the largest bank failure since Washington Mutual in 2008. Its business model was concentrated in a way that made it unusual: it served the technology startup ecosystem, taking deposits from venture-backed companies and investing heavily in long-duration securities during the period of low interest rates. When interest rates rose rapidly through 2022, the market value of those securities fell, creating large unrealized losses; when depositors, concentrated in a single industry and connected through dense professional networks, began to withdraw funds, the bank faced a classic run, accelerated by the speed of digital banking and social media. The bank’s management had failed to hedge its interest rate risk adequately and had not managed its liquidity against the possibility of rapid outflows, failures that would later be documented in detail.

Signature Bank of New York was closed two days later, on March 12, 2023, by the New York Department of Financial Services, which likewise appointed the FDIC as receiver. Signature held roughly 110 billion dollars in assets and had built a significant business serving cryptocurrency firms alongside its traditional commercial real estate lending. Its failure followed a similar dynamic of depositor flight amid the panic triggered by Silicon Valley Bank’s collapse, though its balance sheet problems were distinct. The two failures together, followed within weeks by the failure of First Republic Bank, constituted the most serious episode of banking stress since the 2008 crisis, and they immediately raised the question that the 2018 debate had prefigured: had the re-tiering contributed?

The arithmetic of the question was what made it genuine rather than rhetorical. Both failed banks held assets above the 50 billion dollar line that had triggered automatic enhanced supervision before 2018, and both held assets below the 250 billion dollar line that triggered it after. Under the pre-2018 framework, both would have been subject to the full apparatus of enhanced prudential standards automatically. Under the post-2018 framework, both fell in the discretionary band, where the Federal Reserve could have applied the standards institution by institution but was not required to. That fact alone did not establish causation; many things differ between a counterfactual world and the actual one, and the failures had proximate causes in management decisions and supervisory lapses that the threshold change did not create. But the fact gave the causal question a concrete form that could not be dismissed as speculation. The largest failed bank of the episode sat precisely in the gap the 2018 law had opened, and any honest evaluation of the law had to reckon with that placement.

The federal response to the failures was swift and, by the standards of bank resolutions, extraordinary. The FDIC, the Federal Reserve, and the Treasury Department invoked the systemic risk exception to the least-cost resolution requirement, allowing the FDIC to protect all depositors of both banks, including those with balances above the deposit insurance limit. The Federal Reserve created a new Bank Term Funding Program to provide liquidity to depository institutions against high-quality collateral valued at par. These were crisis measures, taken under authorities that the 2018 statute had left intact, and their necessity underscored the stakes of the causal debate: if the tailoring had contributed to the failures, the lesson was about calibration; if it had not, the lesson was about supervision and management within any plausible calibration.

The Fed’s Own Review: The April 2023 Barr Report

How did the Federal Reserve assess its own supervision after March 2023?

The Federal Reserve assessed its own supervision with unusual candor. Vice Chair for Supervision Michael S. Barr led a review, published April 28, 2023, identifying four contributing factors: the bank’s management failures, supervisors’ failure to appreciate growing vulnerabilities, their failure to act forcefully, and a tailoring framework that impeded effective oversight.

Barr also conceded that stronger requirements might not have prevented the failure, a qualification his critics and his supporters both seized upon.

The review’s four takeaways deserve to be stated in full, because they are the most authoritative factual record of what went wrong and they resist reduction to a single cause. First, the board of directors and management of Silicon Valley Bank failed to manage their risks: the bank grew rapidly without adequate risk management infrastructure, concentrated its exposures, and did not hedge its interest rate risk. Second, Federal Reserve supervisors did not fully appreciate the extent of the bank’s vulnerabilities as it grew in size and complexity; the supervisory ratings and assessments lagged the reality of the bank’s condition. Third, when supervisors did identify vulnerabilities, they did not take sufficient steps to ensure the bank fixed its problems quickly enough; the supervisory process was slow, deferential, and insufficiently forceful. Fourth, and most relevant to the 2018 statute, the Federal Reserve Board’s tailoring approach in response to the 2018 law, combined with a shift in supervisory stance, impeded effective supervision by reducing standards, increasing complexity, and promoting a less assertive supervisory approach.

The fourth takeaway was the one that made headlines, and it requires careful parsing. Barr’s review did not claim that the 2018 statute, by itself, caused the failure. It claimed that the Board’s implementation of the statute’s tailoring mandate, together with a broader shift toward a less assertive supervisory culture, had weakened the supervisory environment in ways that contributed to the outcome. The distinction between the statute and its implementation matters, because the statute’s defenders argued that the discretionary band gave the Federal Reserve all the authority it needed to supervise Silicon Valley Bank rigorously, and that the failure was therefore one of supervision rather than of law. Barr’s review acknowledged the force of that argument even as it criticized the tailoring framework: Barr conceded that higher requirements, applied to the bank, might not have prevented its failure, a concession that complicated any simple causal story.

The review also documented the supervisory history in detail, and the detail cut in multiple directions. Supervisors had identified interest rate risk and liquidity management issues at the bank well before its failure but had pursued them through the slow machinery of matters requiring attention and matters requiring immediate attention, without escalating to enforcement action. The bank’s rapid growth, from tens of billions to over 200 billion dollars in assets in a few years, had outpaced both its own risk management and the supervisory resources assigned to it. These were failures of supervision that could have occurred under any threshold regime, which was the core of the argument that the 2018 law was not the decisive factor. At the same time, the review’s account of how the tailoring framework reduced standards and increased complexity gave the law’s critics a documented, official source for the claim that the re-tiering had weakened the supervisory environment. Both readings of the review were available in its text, and both were taken up in the debate that followed.

The Contested Causal Question

Whether the 2018 re-tiering contributed to the failures of March 2023 is the most consequential unresolved question about the statute, and it must be presented as contested, because it is. Serious analysts, working from the same factual record, have reached opposite conclusions about the causal weight to assign the threshold change, and the honest presentation of the debate requires giving each side its strongest form. What follows names the participants and their arguments as they stood in the dated record, without adjudicating between them.

The case for a tailoring link begins with the Barr review’s fourth takeaway, which is the strongest official evidence the critics possess. A review led by the Federal Reserve’s own Vice Chair for Supervision concluded that the tailoring approach adopted in response to the 2018 statute had impeded effective supervision by reducing standards, increasing complexity, and promoting a less assertive supervisory approach. That is not a claim made by an advocacy group or a partisan; it is the central bank’s assessment of its own conduct, published on April 28, 2023. Senator Elizabeth Warren, who had led the Democratic opposition to the bill in 2018, argued that the failures vindicated her warnings, pointing to the placement of both failed banks in the discretionary band and to the specific relief the law had provided from stress testing and other requirements. Dennis Kelleher of Better Markets contended that the law’s threshold changes had been a decisive deregulatory step, arguing that the automatic application of enhanced standards to banks of Silicon Valley Bank’s size would have constrained the risk-taking that led to its failure. Carter Dougherty of Americans for Financial Reform made a similar case from the advocacy perspective, emphasizing the political economy of the tailoring: that discretionary supervision was inherently more vulnerable to industry pressure and supervisory forbearance than automatic rules. And Professor Rodney Ramcharan of the University of Southern California’s Marshall School of Business offered an academic version of the argument, analyzing the mechanisms by which reduced supervisory intensity could have allowed the banks’ vulnerabilities to grow unchecked.

The logic of the pro-link case can be stated compactly. Before 2018, a bank of Silicon Valley Bank’s size would have faced annual supervisory stress tests, company-run stress tests, enhanced liquidity requirements, and the full apparatus of section 165 automatically. After 2018, it faced a discretionary regime in which the Federal Reserve could apply those requirements but did not apply them all. The stress tests in particular mattered to the argument: a rigorous supervisory stress test, incorporating an interest rate shock scenario, might have forced the bank to confront its unrealized securities losses and its funding fragility before the run began. The counterfactual cannot be tested, but the critics argued that the direction of the effect was clear: less automatic scrutiny meant more room for unmanaged risk to accumulate, and the accumulation was visible in the failed banks’ balance sheets.

The case against a direct link begins with the proposition that the Federal Reserve retained ample authority to supervise the failed banks rigorously and failed to use it, which makes the statute’s threshold change a secondary factor at best. Daniel Tarullo, who had served as the Federal Reserve governor responsible for supervision during the post-crisis years and had been an architect of the original enhanced prudential regime, stated the point directly: “I don’t think there’s that direct of a connection” between the 2018 law and the failures. Tarullo’s argument was that the supervisory tools available to the Federal Reserve, even under the re-tiered framework, were more than sufficient to address the vulnerabilities at Silicon Valley Bank, and that the failure reflected supervisory execution rather than statutory constraint. Thomas Hoenig, the former FDIC vice chairman and a longtime advocate of strong bank regulation, made a related argument: regulators still had the tools they needed and did not use them, which located the failure in the agencies rather than in the law. On this view, the discretionary band was not a gap in authority but a grant of it, and the Federal Reserve’s failure to exercise that authority was the relevant breakdown.

A later contribution to the skeptical side came from within the Federal Reserve itself. In a speech delivered in London in September 2026, Governor Michelle Bowman described preliminary findings from the Starling Advisory Group, an external review of supervision, indicating that the tailoring mandates of the 2018 statute had “did not contribute” to the supervisory delays that preceded the failures. The characterization must carry its qualifications, because they are substantial: the findings were described as preliminary, the full report had not been released at the time of the speech, and Senator Warren publicly contested the conclusions, arguing that they were inconsistent with the Barr review’s documented account of how tailoring had weakened supervision. The Bowman speech therefore belongs in the record as a dated development in the debate, not as its resolution, and the contest over its significance illustrated the broader pattern: every new piece of evidence was absorbed into the existing disagreement rather than settling it.

Weighing the two cases honestly requires acknowledging what each concedes to the other. The pro-link case concedes, through Barr’s own qualification, that stronger requirements might not have prevented the failure; the management failures and the interest rate shock were real independent causes. The skeptical case concedes, through the documented supervisory history, that supervision was in fact less assertive during the tailoring period; the disagreement is about whether the statute caused that softening or merely coincided with a broader cultural shift. The placement of the failed banks in the discretionary band is a fact that favors the critics’ framing, while the Federal Reserve’s unused discretionary authority is a fact that favors the skeptics’. Neither fact determines the causal weighting, which is why the weighting remains contested among informed observers.

The neutrality obligation for this article requires one further observation. The debate over the 2023 failures has been conducted, on both sides, with a selectivity that the record does not support. Critics of the 2018 law sometimes write as though the Barr review established a direct causal chain from the statute to the failures, which it did not; its fourth takeaway was about the tailoring approach and the supervisory stance, and its concession about prevention cut against the strongest version of the claim. Defenders of the law sometimes write as though the supervisory failures documented in the review exonerate the statute entirely, which they do not; a supervisory failure that occurs within a weakened framework is not obviously independent of the framework. The careful reader will hold both the documented facts and the limits of what they establish, and will resist the pressure, considerable in a politically live debate, to convert a contested causal question into a settled one.

The 2023 Reckoning in Congress

The failures of March 2023 returned the 2018 statute to the center of congressional debate, and the hearings that followed replayed the arguments of 2018 with new evidence and higher stakes. The Senate Banking Committee and the House Financial Services Committee held hearings in the spring of 2023 at which regulators, including Vice Chair Barr, testified about the supervision of the failed banks and the role of the tailoring framework. Barr’s testimony tracked his review: management failure first, supervisory failure second, and the tailoring approach as a contributing environmental factor. Senators who had opposed the 2018 bill treated the hearings as vindication, pressing regulators on why discretionary authority had not been used more aggressively and arguing that the episode proved the threshold changes had been a mistake. Senators who had supported the bill treated the hearings as confirmation that supervision, not statute, was the issue, pressing the same regulators on why identified vulnerabilities had not drawn enforcement action when the authority to act was undisputed.

The debate’s structure mirrored the contested causal analysis, because the participants were working from the same record and drawing opposite lessons. For the bill’s critics, the lesson was about the framework: automatic rules had been replaced with discretionary ones, discretion had been exercised weakly, and the weakness was foreseeable because discretionary regimes are inherently more vulnerable to forbearance than automatic ones. The remedy, on this view, was to restore automaticity, to move the lines back down and remove the dependence on supervisory will. For the bill’s defenders, the lesson was about execution: the Federal Reserve had possessed every tool it needed, had identified the banks’ problems, and had failed to escalate, which meant the breakdown would have occurred under any threshold regime. The remedy, on this view, was supervisory reform, not statutory revision, and changing the lines would punish institutions that had done nothing wrong for the failures of supervision at two specific banks.

What the hearings did not produce was a revision of the statute. The 2018 law remained in place through the period this article covers, its thresholds unmoved, its discretionary band intact, its rulemakings operative. That legislative stasis is itself a datum about the amendment stage: once a re-tiering is enacted and implemented, the burden of proof shifts to those who would re-tier again, and a contested causal record does not shift it. The defenders of the 2018 framework could point to the Barr review’s concession and the skeptics’ analyses; the critics could point to the fourth takeaway and the placement of the failed banks. With the evidence balanced as it was, the coalition for change could not assemble the majorities that the coalition for the original bill had commanded. The re-tiering, enacted in calm conditions, survived its first storm not because the storm vindicated it but because the storm’s lessons were genuinely ambiguous, and ambiguity favors the status quo.

Amendment Episodes Compared: The Re-tiering in the Statute’s Longer Life

The 2018 re-tiering was not the first amendment to a major financial statute, and placing it alongside earlier episodes clarifies what was distinctive about it. The pattern of post-enactment revision is as old as federal financial regulation: Congress legislates in crisis, sets thresholds and requirements under time pressure, and then spends the following decade adjusting the calibration as the industry’s complaints and the regulators’ experience accumulate. The Depository Institutions Deregulation and Monetary Control Act of 1980, for example, phased out interest rate ceilings that had become untenable, a substantive change to the rules themselves. The Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994 dismantled geographic restrictions, again changing what banks could do rather than which rules reached them. The Gramm-Leach-Bliley Act of 1999 repealed the Glass-Steagall separation of commercial and investment banking, a substantive repeal of a substantive prohibition. Against this history, the 2018 statute stands out for the purity of its method: it changed almost nothing about what the law required and almost everything about whom the law required it of.

That purity is what makes the episode theoretically interesting and practically confusing. Substantive amendments are easy to evaluate: did the new rule work better than the old one? Coverage amendments are harder, because their effects are distributional rather than directional. The 2018 law did not make the financial system more or less regulated in any aggregate sense that can be measured straightforwardly; it made the regulation more graduated, concentrating intensity at the top of the size distribution and relieving the bottom and middle. Evaluating such a change requires specifying the counterfactual distribution, which requires a theory of where risk resides, which is precisely what the proportionality and automaticity debate was about. The difficulty of the evaluation is not a flaw in the analysis; it is a property of the legislative technique, and it explains why the debate over the 2018 law has been so resistant to resolution.

The episode also illustrates a recurring dynamic of the amendment stage: the coalition that enacts a re-tiering is broader than the coalition that would enact a substantive revision, because re-tiering allows each faction to claim a different victory. In 2018, Republicans claimed regulatory relief, the Democratic negotiators claimed community bank protection, and both claimed to have preserved the core of Dodd-Frank. A substantive revision, a repeal of the Volcker Rule outright, or the abolition of the Bureau, could not have assembled that coalition, because each faction’s victory would have been another’s defeat. The coverage technique’s political genius is that it is positive-sum in its framing: everyone gets relief for their favored institutions, no one loses a favored authority. The technique’s analytical cost is the one this article has traced throughout: when the re-tiered framework is later tested, the positive-sum framing collapses, and each faction discovers that the other’s victory was the source of the vulnerability.

The final comparative observation concerns timing. The 2018 re-tiering arrived eight years after Dodd-Frank’s enactment, at a moment of economic calm and industry profitability, and it was tested five years later under stress. That sequence, calm revision followed by stressed evaluation, is the characteristic rhythm of the amendment stage, and it biases the evaluation in a predictable direction. Revisions made in calm conditions systematically underestimate the value of the protections they trim, because the protections’ value is most visible under stress, which is absent when the revision is designed. The designers of the 2018 law were not unaware of this bias; the discretionary band was, in part, an attempt to correct for it, preserving supervisory capacity for the middle tier against the possibility that calm conditions would not last. Whether the correction was sufficient is the contested question with which this article has been concerned, and the comparison with earlier episodes suggests it is a question every re-tiering will face, because the bias is structural rather than particular to this statute.

Three Recurring Errors in Describing the Statute

The brief for this article identifies three recurring errors in public descriptions of the 2018 law, and each is worth examining because each distorts a different part of the evaluation. The first error is describing the statute as a repeal. The error takes several forms, from the casual headline that calls the law the Dodd-Frank repeal to the sophisticated argument that the threshold changes amounted to a functional repeal for the affected institutions. The casual form is simply false, as the inventory of surviving authorities demonstrates. The sophisticated form deserves a longer answer, because it contains a genuine insight wrapped in a misleading label. It is true that for a bank below 100 billion dollars, the practical difference between the pre-2018 and post-2018 regimes was large: the enhanced prudential standards that had applied automatically ceased to apply at all. But calling that change a repeal confuses the removal of coverage with the removal of the authority, and the confusion has consequences. A repealed authority cannot be restored by supervisory discretion; a re-tiered authority can, because the Federal Reserve retained the power to apply the standards in the discretionary band and Congress retained the power to move the lines again. The repeal framing implies irreversibility and wholesale abandonment, neither of which describes what happened.

The second error is assuming that all banks were relieved. The statute’s benefits were concentrated by design: the Volcker exclusion, the leverage ratio election, the mortgage safe harbor, and the HMDA partial exemption all targeted institutions below the 10 billion dollar line, while the threshold changes benefited the middle tier up to 250 billion dollars. The largest institutions, the money center banks and global trading houses above 250 billion dollars, received no relief at all; every enhanced standard continued to apply to them automatically, and the Volcker Rule, the derivatives clearing mandate, and the resolution planning requirements bound them as before. Even within the relieved tiers, the benefits were conditional rather than blanket: the Volcker exclusion required meeting both the size and trading tests, the leverage ratio required maintaining 9 percent capital, the mortgage safe harbor required portfolio retention and underwriting guardrails. A description of the law as a general bank deregulation misses this targeting, and the miss matters because the targeting was the policy’s rationale. The law’s defenders did not argue that regulation was unnecessary; they argued that it was misallocated, and the allocation is visible in every provision’s qualifying criteria.

The third error is asserting a settled causal link between the 2018 law and the 2023 failures. This error appears in both directions: in claims that the rollback caused the failures, stated as established fact, and in claims that the rollback has been exonerated, stated with equal confidence. The record supports neither certainty. The Barr review documented management failures, supervisory failures, and tailoring-related weaknesses without ranking their causal contributions precisely, and Barr’s concession that stronger requirements might not have prevented the failure explicitly limited what could be claimed. The skeptical analysts documented unused supervisory authority without establishing that its use would have changed the outcome. The preliminary Starling findings described in the September 2026 Bowman speech pointed against a tailoring contribution, but their preliminary status and the public contest over them prevented any settlement. The responsible position, and the one this article maintains, is that the causal weighting is contested: the placement of the failed banks in the discretionary band makes the question genuine, the documented supervisory lapses make it complicated, and the absence of a controlled comparison makes it unresolvable at the level of certainty the public debate often demands.

Four Institutions Under the Re-tiering: A Walkthrough

The re-tiering is easiest to grasp concretely, through the experience of four hypothetical institutions at different points on the size spectrum. Consider first a community bank with 2 billion dollars in total consolidated assets, funded by local deposits and invested in local loans, with negligible trading activity. Before 2018, this bank faced the full risk-based capital framework, the Volcker Rule’s compliance apparatus, the expanded HMDA data fields, and the standard twelve-month examination cycle. After 2018, it could elect the 9 percent Community Bank Leverage Ratio in place of the risk-based framework, it was excluded from the banking entity definition under the Volcker Rule, it was exempt from the expanded HMDA fields if its originations fell below the 500-loan lines, its portfolio loans received the qualified mortgage safe harbor, and it qualified for the eighteen-month examination cycle. For this institution, the statute delivered nearly everything its trade association had requested, and the relief was close to complete.

Consider second a regional bank with 80 billion dollars in total consolidated assets, a traditional commercial lender with a modest trading book. Before 2018, this bank sat above the 50 billion dollar line and faced the full enhanced prudential standards automatically: supervisory and company-run stress tests, risk committee requirements, liquidity obligations, and resolution planning. After 2018, with the law effective immediately for institutions below 100 billion dollars, this bank exited the enhanced regime entirely upon enactment. Its company-run stress tests ended, its resolution planning obligations ended, and its supervision reverted to the ordinary regime. For this institution, the change was the largest in practical terms, because it moved from the full apparatus to none of it in a single day, and it was institutions of this size whose transition the bill’s opponents found most troubling.

Consider third a super-regional bank with 200 billion dollars in total consolidated assets, with significant capital markets activity and cross-border operations. Before 2018, this bank faced the enhanced standards automatically by virtue of exceeding 50 billion dollars. After 2018, it fell in the discretionary band, where the Federal Reserve could apply any or all of the enhanced standards upon determining that doing so would promote financial stability or safety and soundness. In practice, under the tailoring framework finalized in the fall of 2019, such an institution faced a graduated set of requirements reflecting its risk indicators, lighter than the regime above 250 billion dollars but heavier than the regime below 100 billion. For this institution, the statute replaced certainty with conditionality: its regulatory burden depended on supervisory judgment rather than statutory command, which was either the system’s greatest virtue or its greatest vulnerability depending on one’s philosophy.

Consider fourth a money center bank with 2 trillion dollars in total consolidated assets and a global trading operation. Before 2018, this bank faced every Dodd-Frank requirement at maximum intensity. After 2018, nothing changed: the enhanced prudential standards applied automatically above 250 billion dollars, the Volcker Rule applied in full, the derivatives clearing mandate applied, resolution planning applied, and the stress testing regime applied. For this institution, the statute was a nonevent, which was precisely the point. The re-tiering concentrated the post-crisis framework’s intensity where the systemic risk was concentrated and relieved it elsewhere, and the largest institutions’ experience demonstrates that the law’s aggregate regulatory stance, measured at the top of the system, did not move.

The walkthrough reveals the statute’s distributional character more clearly than any summary. The benefits flowed downward along the size spectrum, largest for the smallest institutions, substantial for the middle tier, absent at the top. The risks of the re-tiering, to the extent they materialized, were concentrated in the middle tiers as well, where the transition from automatic to discretionary or unavailable supervision changed the supervisory environment most dramatically. Silicon Valley Bank, at approximately 209 billion dollars, was the third institution in this walkthrough, the super-regional in the discretionary band, and its story is the story of what the band’s conditionality meant in practice when management failed and supervision lagged.

The four cases also clarify why the statute’s defenders and critics could look at the same law and see different things. The defenders looked at the first and fourth institutions, the community bank relieved and the money center untouched, and saw a well-targeted reform. The critics looked at the second and third institutions, the regional bank freed entirely and the super-regional moved to discretion, and saw the removal of automatic safeguards from firms large enough to fail disruptively. Both observations were accurate descriptions of the statute’s effects; the disagreement was about which effects defined the law’s character. That disagreement could not be settled by closer reading of the text, because the text contained both effects by design, and it is still the central unresolved interpretive question about the 2018 law.

The Tier Table

The artifact for this article is the tier table: each major requirement, the threshold that applied before 2018, the threshold that applied after, and whether application became automatic, discretionary, or unavailable. The table is the re-tiering made visible, and it rewards close reading. Notice how few of the rows show a requirement disappearing; most show it moving, narrowing, or becoming conditional. Notice also the pattern in the “after” column: the higher the systemic stakes of the requirement, the more likely it was to survive in automatic or discretionary form for the largest institutions, while the requirements most burdensome for small institutions were the ones most likely to become elective or unavailable. The table is the statute’s theory of itself, rendered as data.

Requirement	Threshold before 2018	Threshold after 2018	Application after 2018
Enhanced prudential standards, Dodd-Frank section 165	Automatic at 50 billion dollars or more in total consolidated assets	Automatic at 250 billion dollars or more; Federal Reserve discretion for 100 to 250 billion dollars; exemption below 100 billion dollars	Automatic above 250 billion; discretionary from 100 to 250 billion; unavailable below 100 billion
Supervisory stress tests	Annual for bank holding companies at or above 50 billion dollars	Annual for bank holding companies at or above 250 billion dollars; discretionary for 100 to 250 billion dollars	Automatic above 250 billion; discretionary from 100 to 250 billion; unavailable below 100 billion
Company-run stress tests	Required at or above 10 billion dollars in total consolidated assets	Required at or above 250 billion dollars in total consolidated assets	Automatic above 250 billion; unavailable below 250 billion
Mandatory risk committee, publicly traded bank holding companies	Required at or above 10 billion dollars in total consolidated assets	Required at or above 50 billion dollars in total consolidated assets	Automatic above 50 billion; unavailable below 50 billion
Volcker Rule, proprietary trading and covered funds	Applied to all banking entities	Institutions at or below 10 billion dollars with trading assets and liabilities at or below 5 percent of total assets excluded from the banking entity definition	Unavailable for qualifying institutions; automatic for all others
Community Bank Leverage Ratio	Not available; risk-based capital framework applied to all	Elective 9 percent Tier 1 leverage ratio for qualifying institutions below 10 billion dollars in total consolidated assets	Elective for qualifying institutions; unavailable above the qualifying criteria
Qualified mortgage safe harbor, portfolio loans	Available under the general ability-to-repay framework	Extended to loans originated and retained by institutions below 10 billion dollars meeting the statutory guardrails	Elective for qualifying originators; general framework unchanged for others
HMDA expanded data fields, 2015 rule	Reportable by all covered lenders	Partial exemption for insured depositories and credit unions below 500 originations in each of the two preceding calendar years; 22 pre-existing fields still required	Unavailable as an obligation for qualifying institutions; automatic for all others
Supplementary leverage ratio, custody banks	Central bank deposits included in leverage exposure	Central bank deposits excluded for qualifying custody banks	Adjusted automatically for qualifying custody banks; unchanged for others
Extended examination cycle	Available to qualifying institutions at or below 1 billion dollars in total assets	Available to qualifying institutions at or below 3 billion dollars in total assets	Elective for qualifying institutions up to the new line; unchanged above it
Municipal obligations as high-quality liquid assets	Qualifying municipal obligations not treated as Level 2B	Investment grade, liquid, readily marketable municipal obligations treated as Level 2B	Automatic for qualifying securities; unchanged for others
Resolution planning, Dodd-Frank section 165(d)	Required at or above 50 billion dollars in total consolidated assets	Required at or above 250 billion dollars in total consolidated assets	Automatic above 250 billion; unavailable below 250 billion
Orderly Liquidation Authority, Title II	Available for systemically important financial companies	Unchanged	Unchanged and available
FSOC designation authority	Available for nonbank financial companies	Unchanged	Unchanged and available
CFPB structure, funding, and authorities	As established by Dodd-Frank	Unchanged	Unchanged and available
Derivatives clearing mandate, Title VII	As established by Dodd-Frank	Unchanged	Unchanged and available

Studying the Re-Tiering

The re-tiering is best studied as a system rather than as a list of provisions, and the study repays a particular discipline: for each provision, identify the requirement, the institutions it reached before 2018, the institutions it reached after, and the mechanism, automatic, discretionary, elective, or unavailable, by which it reached them. Readers who work through the statute in that order will find that the political arguments about it become easier to evaluate, because most of those arguments turn out to be arguments about a specific row of the tier table rather than about the statute as a whole. The VaultBook legislation study notebook provides a structured template for that kind of provision-by-provision analysis, with fields for the statutory citation, the pre- and post-2018 thresholds, and the application mechanism.

The Amendment Stage, Concluded

The Economic Growth, Regulatory Relief, and Consumer Protection Act was the Dodd-Frank Act’s first great amendment, and it established the pattern that the amendment stage would follow: not the repeal that the most ardent critics of the 2010 law wanted, and not the untouched preservation that its most ardent defenders wanted, but a recalibration of coverage, negotiated across party lines, that left the architecture standing and moved its walls. The statute’s supporters got relief for the institutions they considered overburdened; its opponents got the preservation of every major authority and a discretionary band that kept the largest regional banks within the Federal Reserve’s reach. Both sides got something, and both sides would later argue that what the other side got was the problem.

The honest assessment, with the full record in view, is that the re-tiering was neither the repeal its opponents described nor the harmless technical correction its most enthusiastic supporters implied. It was a genuine change in the supervisory regime for a defined set of institutions, made on the basis of a plausible but contestable judgment about where systemic risk began, and its consequences included both the intended relief for small lenders and the unintended contribution, disputed in its magnitude but documented in its mechanism, to the supervisory environment in which the 2023 failures occurred. The amendment stage of a statute’s life does not produce verdicts; it produces new arrangements, which the next decade then tests. The test arrived in March 2023, and the grading of it continues.

Frequently Asked Questions

Q: What did the 2018 Dodd-Frank rollback change?

The 2018 statute, Public Law 115-174, changed which institutions faced Dodd-Frank’s toughest requirements rather than changing the requirements themselves. Its central move raised the asset threshold for automatic enhanced prudential standards from 50 billion to 250 billion dollars in total consolidated assets, with a discretionary band from 100 to 250 billion dollars where the Federal Reserve could apply standards institution by institution. It excluded small, low-trading banks from the Volcker Rule’s banking entity definition, created an elective 9 percent Community Bank Leverage Ratio for qualifying institutions under 10 billion dollars, extended a qualified mortgage safe harbor to portfolio loans at small lenders, and granted partial HMDA reporting relief. Targeted provisions addressed custody banks, examination cycles, and municipal securities. No major Dodd-Frank authority was repealed.

Q: What is the Dodd-Frank SIFI threshold now?

There is no single SIFI threshold in the re-tiered framework; the answer depends on the requirement. Enhanced prudential standards under Dodd-Frank section 165 apply automatically at 250 billion dollars or more in total consolidated assets, with Federal Reserve discretion to apply them to institutions between 100 and 250 billion dollars where doing so would promote financial stability or safety and soundness. Institutions below 100 billion dollars were exempted immediately upon enactment in May 2018. Related thresholds moved as well: company-run stress tests rose from 10 to 250 billion dollars, mandatory risk committees for publicly traded bank holding companies rose from 10 to 50 billion dollars, and resolution planning under section 165(d) follows the 250 billion dollar line. The old uniform 50 billion dollar trigger no longer governs any of these requirements.

Q: Which banks were exempted from the Volcker Rule in the Dodd-Frank rollback?

Banks meeting two conjunctive tests were excluded from the Volcker Rule: total consolidated assets at or below 10 billion dollars, and trading assets and liabilities at or below 5 percent of total consolidated assets. The mechanism is frequently misdescribed. Section 203 did not merely waive the proprietary trading ban for these institutions; it removed them from the definition of banking entity under section 13 of the Bank Holding Company Act, so the entire Volcker Rule, both the proprietary trading prohibition and the covered fund restrictions on hedge fund and private equity investments, ceased to apply. An institution had to satisfy both the size test and the trading activity test. Banks above 10 billion dollars, and smaller banks with outsized trading books, remained fully subject to the rule.

Q: Was the 2018 Dodd-Frank rollback bipartisan?

Yes, by the standards of financial legislation in that decade. The Senate passed S. 2155 on March 14, 2018, by 67 to 31, with all Republicans present joined by sixteen Democrats and independent Angus King of Maine, who caucused with the Democrats. The House passed it on May 22, 2018, by 258 to 159, with all Republicans except Walter B. Jones Jr. of North Carolina joined by 33 Democrats. The bipartisanship was structural, not decorative: with 51 Republicans in the Senate, the bill needed Democratic votes to reach the sixty required for cloture, and it was negotiated from the start to attract a bloc of Democratic senators from states with large community bank constituencies. Prominent Democratic opponents included Elizabeth Warren, Chuck Schumer, and Sherrod Brown, and the party split publicly over the bill.

Q: Did the Dodd-Frank rollback contribute to bank failures?

The causal weighting is contested among informed observers, and the honest answer reflects that. Silicon Valley Bank, closed March 10, 2023, with about 209 billion dollars in assets, and Signature Bank, closed March 12, 2023, with about 110 billion dollars, both sat above the old 50 billion dollar automatic threshold and below the new 250 billion dollar one, placing them in the discretionary band the 2018 law created. The Federal Reserve’s April 28, 2023 review, led by Vice Chair Michael Barr, found that the tailoring approach had impeded effective supervision, though Barr conceded stronger requirements might not have prevented the failure. Critics including Elizabeth Warren and analysts at Better Markets and Americans for Financial Reform argued the link was real; skeptics including Daniel Tarullo and Thomas Hoenig argued regulators retained adequate tools and failed to use them.

Q: What is the community bank leverage ratio in the Dodd-Frank rollback?

The Community Bank Leverage Ratio, created by section 201, is an elective simplified capital measure for qualifying small institutions. A bank that maintains the ratio is deemed to satisfy the generally applicable risk-based capital requirements without computing risk-weighted assets. The statute directed the banking agencies to set the ratio between 8 and 10 percent; the final rule of October 29, 2019, set it at 9 percent of Tier 1 capital to average total consolidated assets. Qualifying institutions must hold less than 10 billion dollars in total consolidated assets, keep off-balance-sheet exposures at or below 25 percent of average total consolidated assets, keep trading assets and liabilities at or below 5 percent, and not be advanced-approaches organizations. A two-quarter grace period applies if the ratio slips to between 8 and 9 percent.

Q: Was Dodd-Frank repealed?

No. The 2018 statute repealed no major Dodd-Frank authority. The Orderly Liquidation Authority, the Financial Stability Oversight Council’s designation power, the Consumer Financial Protection Bureau’s structure, funding, and authorities, the Title VII derivatives clearing mandate, the Volcker Rule for banks above the small-bank exclusion line, and resolution planning for firms above 250 billion dollars all survived intact. What changed was coverage: the set of institutions to which each requirement applied, and whether application was automatic, discretionary, elective, or unavailable. The distinction matters because the two most common public arguments about the law, that Dodd-Frank was repealed and that the tailoring was miscalibrated, are entirely different claims. The first is false as a matter of statutory text; the second is the subject of genuine and continuing dispute.

Q: Which Democrats voted for the Dodd-Frank rollback?

Sixteen Democratic senators voted yes on S. 2155 in the Senate’s March 14, 2018 roll call: Michael Bennet of Colorado, Tom Carper of Delaware, Chris Coons of Delaware, Joe Donnelly of Indiana, Maggie Hassan of New Hampshire, Heidi Heitkamp of North Dakota, Doug Jones of Alabama, Tim Kaine of Virginia, Joe Manchin of West Virginia, Claire McCaskill of Missouri, Bill Nelson of Florida, Gary Peters of Michigan, Jeanne Shaheen of New Hampshire, Debbie Stabenow of Michigan, Jon Tester of Montana, and Mark Warner of Virginia. Independent Angus King of Maine, who caucused with the Democrats, also voted yes, bringing the Democratic caucus supporters to seventeen members in total. Thirty-one Democrats and independent Bernie Sanders of Vermont voted no. In the House, 33 Democrats voted for the bill on May 22, 2018.

Q: When did the 2018 statute’s main provisions take effect?

The statute used staggered effective dates. Bank holding companies below 100 billion dollars in total consolidated assets were freed from the enhanced prudential standards immediately upon enactment on May 24, 2018. The remaining amendments, including the new 250 billion dollar automatic threshold and the Federal Reserve’s discretionary authority over the 100 to 250 billion dollar band, took effect eighteen months after enactment, on November 24, 2019, giving regulators time to write implementing rules. The Volcker Rule exclusion for qualifying small banks took effect immediately. The Community Bank Leverage Ratio required agency rulemaking; the final rule setting it at 9 percent was issued October 29, 2019. The custody bank leverage provision’s final rule took effect April 1, 2020, and the extended examination cycle rule took effect January 28, 2019.

Q: What happened to stress tests for mid-sized banks under the 2018 statute?

The threshold for mandatory company-run stress tests rose from 10 billion to 250 billion dollars in total consolidated assets, eliminating the requirement for a large group of mid-sized institutions. Supervisory stress tests conducted by the Federal Reserve, previously annual for bank holding companies at or above 50 billion dollars, moved to the new framework: automatic at or above 250 billion dollars, with Federal Reserve discretion to apply them to institutions between 100 and 250 billion dollars. Banks below 100 billion dollars were no longer subject to either form of stress testing under the enhanced standards. The change concentrated the testing regime on the largest firms and made testing conditional rather than automatic for the middle tier, which became one of the most debated features of the re-tiering after the 2023 failures.

Q: What changed for mortgage lending at small banks in the 2018 statute?

Section 101 created a qualified mortgage safe harbor for residential mortgage loans originated and retained in portfolio by depository institutions and credit unions with less than 10 billion dollars in total consolidated assets. A qualified mortgage gives the lender a safe harbor or presumption of compliance with the ability-to-repay requirement, a significant legal protection. Qualifying loans had to avoid negative amortization and interest-only features, observe limits on points and fees, and reflect a documented process in which the lender considered and verified the borrower’s debts, income, and financial resources. The safe harbor generally lasted only while the originating institution held the loan, with transfer allowed in narrow cases. The portfolio-retention condition tied the legal benefit to the economic incentive: a lender holding its own loans bears the loss on bad underwriting.

Q: How did the 2018 statute change HMDA reporting for small lenders?

Section 104(a) gave partial relief from the Home Mortgage Disclosure Act’s expanded data fields. Insured depository institutions and credit unions that originated fewer than 500 closed-end mortgages or 500 open-end lines of credit in each of the two preceding calendar years were exempted from reporting the additional data fields added by the Consumer Financial Protection Bureau’s 2015 rule. The exemption was partial: the 22 data points lenders had reported before the 2015 expansion remained fully required. It covered only depositories and credit unions, excluding nonbank lenders, and excluded institutions with poor Community Reinvestment Act ratings. Contrary to some summaries, the provision contained no two-year sunset, and it did not eliminate mortgage data reporting for qualifying lenders.

Q: What did the 2018 statute do for custody banks?

Section 402 excluded central bank deposits from the supplementary leverage ratio for qualifying custody banks. Custody banks safekeep assets for institutional clients, and client cash awaiting investment sits as deposits that the custody bank places at the central bank. Before 2018, those deposits counted as leverage exposure under the supplementary leverage ratio, generating a capital requirement against assets that were among the safest on any balance sheet. The provision aligned the capital treatment with the economic reality of the custody business model. The Federal Reserve’s final rule implementing section 402 was issued November 19, 2019, and took effect April 1, 2020. The change was narrow and technical, and it illustrated the statute’s tailoring method at its most precise: adjusting a rule to fit a specific business model rather than repealing the rule.

Q: What consumer protections did the 2018 statute add?

The statute added consumer-facing protections even as it reduced regulatory burdens elsewhere. Section 309 imposed a seasoning requirement on certain refinancing loans guaranteed by the Department of Veterans Affairs, responding to rapid serial refinancings that stripped equity from veteran borrowers through repeated fees. Other provisions addressed private student loan debt after the death or total and permanent disability of the borrower, requiring lenders to release cosigners and discharge obligations in circumstances that had generated sustained complaints. The statute also assigned the Consumer Financial Protection Bureau new implementation work on its mortgage and reporting provisions. These additions complicated the one-directional deregulation narrative: the 2018 Congress paired relief for small institutions with new protections aimed at specific consumer harms it had identified.

Q: Why did some Democrats oppose the 2018 statute while others supported it?

The Democratic split turned on how far up the size spectrum the relief should reach. Supporters such as Jon Tester, Heidi Heitkamp, Joe Donnelly, and Mark Warner argued that Dodd-Frank’s fixed compliance costs fell hardest on community banks that posed no systemic risk, that the burdens were driving consolidation and reducing credit in their states, and that tailoring supervision to risk was sound policy. Opponents led by Elizabeth Warren, joined by Chuck Schumer, Sherrod Brown, and advocacy groups including Americans for Financial Reform and Better Markets, argued that the threshold changes extended far beyond genuinely small institutions, that the discretionary band would leave large regional banks under-supervised, and that the community bank framing provided political cover for relief flowing mainly to much larger firms. Both factions agreed small lenders deserved help; they disagreed about the calibration.

Q: What did the April 2023 Federal Reserve review conclude about Silicon Valley Bank?

The Review of the Federal Reserve’s Supervision and Regulation of Silicon Valley Bank, led by Vice Chair for Supervision Michael S. Barr and published April 28, 2023, identified four contributing factors. First, the bank’s board and management failed to manage its risks, growing rapidly without adequate controls and leaving interest rate risk unhedged. Second, Federal Reserve supervisors did not fully appreciate the extent of the bank’s vulnerabilities as it grew in size and complexity. Third, when supervisors identified problems, they did not act quickly or forcefully enough to ensure correction. Fourth, the Board’s tailoring approach in response to the 2018 statute, combined with a shift in supervisory stance, impeded effective supervision by reducing standards, increasing complexity, and promoting a less assertive supervisory approach. Barr also conceded that higher requirements might not have prevented the failure.

Q: Who argued that the 2018 statute did not cause the 2023 failures, and what was their reasoning?

Daniel Tarullo, the former Federal Reserve governor who had overseen supervision during the post-crisis years, said he did not see a direct connection, arguing that the supervisory tools available even under the re-tiered framework were sufficient and that the breakdown was one of execution rather than authority. Thomas Hoenig, the former FDIC vice chairman, argued along similar lines that regulators retained the tools they needed and failed to use them. In a September 2026 speech in London, Governor Michelle Bowman described preliminary findings from the Starling Advisory Group suggesting the tailoring mandates had not contributed to supervisory delays. Those findings were preliminary, the full report was unreleased, and Senator Warren contested them. The common thread of the skeptical case is that discretionary authority was a grant of power, not a gap, and the failure lay in its exercise.

Q: What is the difference between the supplementary leverage ratio change and the community bank leverage ratio?

The two provisions share the word leverage but address different rules for different institutions. The supplementary leverage ratio change in section 402 was a definitional adjustment for custody banks: it excluded their central bank deposits from the exposure measure used in the supplementary leverage ratio, reducing a capital requirement that their business model generated against essentially risk-free assets. The Community Bank Leverage Ratio in section 201 was an elective alternative capital framework for qualifying institutions under 10 billion dollars: a bank maintaining the 9 percent ratio was deemed to satisfy the risk-based capital requirements without computing risk-weighted assets. One was a targeted fix to an existing ratio’s inputs; the other was an optional substitute for an entire capital regime. Neither repealed the underlying capital rules.

Q: Did the 2018 statute change the Volcker Rule for large banks?

No. The Volcker Rule continued to apply in full to every banking entity above the section 203 exclusion line. Banks with more than 10 billion dollars in total consolidated assets, and smaller banks whose trading assets and liabilities exceeded 5 percent of total assets, remained subject to both the proprietary trading prohibition and the covered fund restrictions on hedge fund and private equity investments. The implementing regulations produced by years of interagency work continued to govern these institutions unchanged. The 2018 change reached only institutions that were both small and minimally engaged in trading, removing them from the banking entity definition entirely. The largest trading houses, whose activities had motivated the rule’s creation, were untouched, which is why the provision is best understood as a coverage change rather than a substantive revision of the rule.

Q: What did the 2018 statute leave unchanged at the Consumer Financial Protection Bureau?

Everything structural. The 2018 statute did not alter the Bureau’s organization, its independent funding through the Federal Reserve, its rulemaking authority, its supervision and examination powers, or its enforcement jurisdiction. The single-director leadership structure, the funding mechanism outside the appropriations process, and the scope of the Bureau’s authority over consumer financial products all survived exactly as the Dodd-Frank Act had established them. What the statute did was give the Bureau additional work: implementing the qualified mortgage safe harbor, the partial HMDA exemption, and the new consumer protections on VA refinancings and student loan discharges. The contrast between the heated political rhetoric about the Bureau during the bill’s debate and the statute’s actual treatment of it remains one of the clearest illustrations of the re-tiering principle.