Stephen Miran addressed the Economic Club of Miami on March 26, 2026, not as an outside commentator surveying Federal Reserve policy from a distance but as a sitting Governor with an FOMC vote and a documented institutional record. Before his confirmation to the Board, he co-authored a staff working paper that laid out the operational blueprint for shrinking the Fed's balance sheet — the same architecture he previewed that evening in Miami.
The position is unusual. Most public arguments for balance sheet reduction originate outside the institution — from academic economists, from former officials, from market participants with an interest in the outcome. Miran's argument originates from inside. He helped design the proposal while working adjacent to the institution, brought it with him when he joined the Board, and is now advancing it in his capacity as one of the seven Governors who will vote on whether and how it is implemented.
The speech is technical and detailed. It names several mechanisms, framed explicitly as a sample rather than a closed list. Among the operational tools aimed at reducing reserve demand, Miran highlights destigmatizing the discount window so banks can hold fewer reserves and easing Liquidity Coverage Ratio requirements to reduce the regulatory demand for those reserves. Separately, he advances a mandate argument: that normalizing the Fed's mortgage-backed securities holdings would remove what he characterizes as an unwanted distortion in credit allocation — a compositionally distinct claim about what the balance sheet should hold, not only how large it should be. He also names additional measures this post does not evaluate in full — among them, more active open market operations around quarter-ends and fiscal settlement dates, standing repo facility operations, and daylight overdraft policy — several of which target the kind of distributional friction that produced the September 2019 repo stress. This post tests the two operational reserve-demand mechanisms and the MBS mandate argument against the documentary record. The remaining measures are noted but not adjudicated here.
Each of these mechanisms rests on a factual premise: that the discount window can be made to function reliably, that the LCR constraint can be eased without replacing the insurance it provides, that credit allocation through MBS holdings can be unwound cleanly, and that reserves can fall to a new floor without repeating the stress of 2019. The Federal Reserve's own documentary record — transcripts, staff memos, supervisory guidance, and the statements of officials who built and monitored this architecture — speaks directly to each premise.
These are testable claims.
We went into the archives.
The Architecture They Chose
Miran's foundational premise holds that the Fed's enlarged balance sheet is not a choice but a consequence — the accumulated result of Dodd-Frank requirements, Basel III liquidity standards, and the altered market structures they produced. "The growth in currency demand, the post-crisis regime put in place by the Dodd–Frank Act and reforms to the Basel standards, and the resulting changes to market structures and expectations all resulted in greater demand for reserves in the system." The passive construction is telling. Reserves accumulated because the environment demanded them. What the 2018–2019 FOMC framework deliberations actually show is something different: an institution that examined its alternatives, weighed the evidence, and chose.
The committee's internal analysis, circulated in staff memos throughout late 2018, compared a floor system against a corridor system — the pre-crisis model in which the Fed controlled short-term rates through daily open market operations against scarce reserves. Staff identified that returning to the corridor would require forecasting reserve demand with a precision the institution acknowledged it could no longer achieve under post-crisis regulatory conditions; staff judged the corridor regime would likely require large, frequent daily operations and more precise reserve-demand forecasting than the institution could achieve under those conditions. The corridor was not merely less convenient — it faced significant operational barriers under the post-crisis regulatory environment. Publicly, however, the January 2019 FOMC statement framed the ample-reserves framework as a preference — a design choice grounded in efficiency and simplicity — while staff analysis had separately concluded that the corridor alternative faced significant operational barriers under post-crisis conditions, a gap between internal assessment and public framing that the January 2019 statement did not surface.
That internal judgment shaped how Chairman Powell characterized the deliberation in November 2018:
"The current system has clear advantages over a system in which reserves are scarce and in which the funds rate is adjusted using frequent open market operations. At a minimum, the scarce-reserves system raises many questions."
By January 2019, the assessment had become institutional consensus. Atlanta Fed President Raphael Bostic made the affirmative case directly:
"the risks of going home to a framework featuring a limited reserves requirement are significant, and an abundant-reserves regime is more robust and agile in the face of financial and economic distress."
The endorsements extended to the official whose public statements come closest to validating Miran's causal framing. In a February 2019 speech, Vice Chair for Supervision Randal Quarles acknowledged that reserve demand had grown "in part" because of "regulatory changes introduced after the crisis," explicitly naming the Liquidity Coverage Ratio: banks were holding reserves to satisfy LCR requirements, and because the regulatory changes and quantitative easing had arrived simultaneously, the underlying demand curve had become, by Quarles's own assessment, nearly impossible to read in isolation. His analysis was the most direct institutional acknowledgment that post-crisis regulation had structurally reshaped reserve demand — and it was the strongest internal voice available for Miran's diagnostic framing.
And yet Quarles never argued for returning to the corridor. He traced the LCR mechanism with precision, named regulation as a structural driver, and endorsed the framework regardless. That endorsement from the institution's own regulatory hawk substantially weakens the involuntary-accommodation reading: post-crisis regulation materially constrained the available options, and within that constrained environment the FOMC affirmatively preferred the floor system on operational grounds rather than merely acquiescing to it. What demands examination is the mechanism that made the corridor operationally inferior under post-crisis conditions, and why the floor system's displacement of it was, by 2019, already structural rather than provisional.
The Window Nobody Opens
The floor system persists, and the balance sheet with it, because the backstop channel that would allow banks to hold fewer reserves has never functioned reliably enough to serve as a substitute. The discount window — the Federal Reserve's lender-of-last-resort facility — should in theory make large reserve buffers unnecessary: banks that can access emergency liquidity on demand need not hold it in advance. Miran's reform architecture rests entirely on this substitution logic. His proposal to "destigmatize the standing repo operations, discount window usage, and daylight overdraft usage" treats stigma as an operational friction — a problem of procedure, familiarity, and regulatory incentive that administrative reform can solve. The archive's verdict on that assumption spans more than two decades and four distinct reform episodes.
The institution has not been passive. The 2003 restructuring created the primary credit facility specifically to reduce stigma by removing the requirement that banks exhaust private credit before approaching the window. The 2007–2008 Term Auction Facility was built explicitly as a stigma workaround — a mechanism to deliver emergency liquidity without the reputational signal that direct discount window borrowing carried. The 2020 pandemic interventions included public encouragement to borrow, changes to the H.4.1 statistical release designed to make large borrowings harder to identify by district, and explicit framing of window use as a sign of institutional health. The 2021 Standing Repo Facility offered yet another channel designed to function as a backstop without the window's reputational weight. Each episode was publicly described as progress.
Internally, the institution reached a different conclusion. In 2016, after thirteen years of post-restructuring evidence, a dedicated working group delivered its verdict:
"The DWWG concluded that it may be very difficult to reduce the stigma associated with primary credit in its current form. One reason for this difficulty is that stigma is inherent to the facility's backstop role."
Inherent to the role. Not a deficiency of design, not a failure of collateral terms, not a problem of communication. The signal that window borrowing sends — that a bank requires emergency access — cannot be separated from the facility's function as an emergency backstop. Jeffrey Lacker had identified the precise mechanism eight years earlier:
"We can try to keep it secret, but there's a broad ability in the market to infer when somebody goes to the window by their behavior before that. I'm saying that I don't think we should get our hopes up about ever eliminating stigma."
Stigma does not require disclosure. Markets infer window use from observable behavior — funding patterns, balance sheet movements, market positioning — regardless of what reporting rules govern disclosure. Neel Kashkari, surveying the same landscape in March 2020, concluded that hoping administrative measures would eliminate stigma so the TAF could be retired was, in his assessment, "pretty optimistic." The internal/public gap in this domain is among the most severe in the archive: staff classified stigma as a permanent informational equilibrium while leadership publicly described each successive reform as progress toward a functional window. Those two registers never converged.
Sixteen years after Lacker's warning, Michelle Bowman — a sitting Governor and Miran's colleague on the Board — placed the institution's current position on record: the Federal Reserve "cannot entirely eliminate discount window borrowing stigma through regulatory fiat," because the stigma derives not from rules but from the fundamental ambiguity of why any given institution is borrowing at all.
The post-2019 record offers partial qualification but not reversal. The 2020 H.4.1 reporting changes — designed to make district-level borrowing harder to identify — along with public official encouragement during the pandemic produced a documented, if temporary, increase in window usage. Philip Jefferson, in 2023, assessed those reforms as meaningful steps forward. The 2021 Standing Repo Facility occupies analytically different ground: internal deliberations, including Lorie Logan's October 2019 assessment and Eric Rosengren's June 2019 framing, treated the SRF explicitly as a mechanism that "might not suffer from the stigma attached to discount window borrowing" by presenting it as a routine monetary policy tool for rate control rather than an emergency backstop. Daylight overdraft access carries a still-different reputational profile, oriented toward intraday settlement rather than stress-liquidity signaling. Miran names all three in the same breath; the archive does not treat them identically.
The qualification has a ceiling, however. The 2016 working group analyzed the same disaggregation and warned that engineering one facility to appear stigma-free may "eliminate stigma at the monetary policy facility — but it may do so by transferring additional stigma to facilities that are seen as liquidity backstops." Neel Kashkari's 2019 assessment was blunter: the Fed is "the source of our stigma," and technical rebranding cannot resolve that if the institution continues to communicate that any facility is a backstop of last resort. The SRF question is less settled than the discount window question; but whether the SRF's routine-use design holds under genuine stress — the condition that matters for reserve substitution — has not been tested at scale. What Bowman's conclusion establishes is a persistent structural limitation on the discount window specifically: a proposal that depends on the window becoming a reliable substitute for precautionary reserve holdings inherits that limitation directly, and the ample reserves built to compensate for the window's unreliability cannot be reduced by the very facility whose failure made them necessary.
The Floor They Already Hit
If the discount window cannot serve as a functional backstop — a conclusion the institution's own working group and sitting Governors have reached — then the LCR is not a regulatory artifact that inflated reserve demand artificially. It is the deliberate engineering of a substitute system, a mechanism that compels banks to hold their own insurance precisely because they will not approach the window under stress. Miran's proposal of "easing liquidity coverage ratio (and related) requirements" assumes this substitution can be undone without consequence. The archive documents the calculation that created it, and then supplies the empirical test: the balance sheet level Miran now targets as achievable is the level at which the system failed in September 2019.
The first finding — the engineering rationale — was stated with unusual directness by Governor Jeremy Stein in April 2013, as the post-crisis liquidity framework was being assembled:
"The introduction of liquidity regulation after the crisis can be thought of as reflecting a desire to reduce dependence on the central bank as a lender of last resort, based on the lessons learned over the previous several years."
The lessons learned were lessons about stigma. Staff analysis from the following years made the mechanism explicit: the LCR required banks to self-insure against potential funding difficulties, holding enough high-quality liquid assets to survive thirty days of stress without approaching the window. Reserves counted toward this buffer. Reserve demand was therefore not an incidental byproduct of the regulatory framework — it was the framework's operational spine. A 2017 staff memo noted plainly that regulatory and other structural factors may now be contributing to higher levels of reserve demand even in more normal times. Easing the LCR removes the workaround. It does not repair the structural failure that made the workaround necessary.
The second finding arrives through the empirical record rather than the regulatory design documents. Miran cites "18 percent of GDP" as his achievable target, noting that the balance sheet occupied roughly that level in 2019 before the pandemic. What the archive records from that year is a system that broke at precisely that level.
Board staff had flagged the structural problem a year earlier. In October 2018 outreach, many banks stated that they do not view reserves and other HQLA as perfect substitutes, and pointed to factors beyond the LCR and relative return as important determinants of demand for reserves. The Senior Financial Officer Survey — the instrument the Committee used to estimate the reserve demand floor — could not capture the distributional frictions that would prevent banks from lending reserves into stress markets even when aggregate levels appeared sufficient. The analytical gap was identified internally. It did not materially alter the normalization trajectory.
September 2019 supplied the correction. At the October 30 meeting convened after the disruption, Governor Lael Brainard stated that reserves had clearly fallen below demand, that Treasury security issuance was rising to historically elevated levels, and that there were frictions in the pipes. That diagnosis stood in direct contrast to the Committee's public presentation of normalization as an orderly, well-calibrated process — a framing maintained throughout the preceding years of balance sheet runoff. The gap between those two registers was not ambiguous.
The Chair's own assessment at that meeting closed the interpretive space:
"Second, reserves are below the desired level and need to be increased. We've got a good plan for that, and we'll learn much more about market functioning as we raise the level of reserves. Third, frequent operations, even if small, are, by definition, not consistent with ample reserves."
The Chair of the Federal Reserve declared that reserves were insufficient at the benchmark Miran now presents as a viable destination. The window is broken, the insurance was built to compensate for it, and the institution's most recent experience at comparable balance-sheet levels ended in stress that Miran's proposed reforms — destigmatization, LCR easing, and distributional-friction remedies — have not yet been demonstrated to prevent. Brainard's diagnosis at that October 30 meeting named not a single cause but a confluence: reserves had fallen below demand, Treasury issuance was rising to historically elevated levels, and there were frictions in the pipes that prevented even nominally sufficient aggregate reserves from reaching the institutions that needed them. Whether Miran's additional operational measures — more active open market operations, standing repo facility deployment, daylight overdraft policy — would have altered that outcome under those conditions is a question the record cannot resolve, because the reformed regime was never tried. What the record does establish is the evidentiary burden that falls on the proposal: to demonstrate that the confluence of conditions that produced 2019 would be materially different under the reformed architecture, before reserves are reduced to levels where the answer becomes consequential.
The Channel They Chose
Of all the claims Miran advances, the one about MBS holdings and credit allocation commands the widest assent. By purchasing and retaining mortgage-backed securities at scale, "the Fed preferentially injects credit into the housing sector in ways it does not for other sectors of the economy" — and in doing so crosses from monetary into fiscal territory, exercising sector-specific credit allocation that an unelected central bank was never authorized to perform. The instinct that this is wrong is broadly distributed across hawks, doves, and institutional observers of every persuasion. The balance sheet's scale carries documented costs in operational flexibility. Its composition carries a different cost — one measured in mandate rather than reserves — and the credit-allocation diagnosis was debated inside the FOMC in real time, advanced by multiple members, and overridden.
Jeffrey Lacker, Richard Fisher, Charles Plosser, Thomas Hoenig, and Esther George each raised, across multiple meetings, the concern that purchasing agency MBS constituted sector-specific credit allocation — a fiscal function the committee was neither equipped nor authorized to perform. The objection was not peripheral. It was a sustained institutional position held by some of the committee's most experienced members, advanced at the moment the program was being designed. Miran's diagnosis is their diagnosis, updated by fifteen years.
The committee heard the objection and chose MBS anyway. The staff analysis explains why.
"Because private interest rates have a greater direct effect on spending than Treasury rates, programs that drive down private rates more than Treasury rates are estimated to have a more potent effect on economic activity than equal-sized programs that lower Treasury rates more than private rates."
The key word is "potent." The staff was not describing an incidental effect on the mortgage market. It was comparing purchase strategies and recommending MBS because the housing credit channel produced stronger stimulus than equivalent Treasury purchases. The sectoral preference was not a distortion alongside the transmission mechanism. It was the transmission mechanism.
The formal FOMC directive from August 2009 named the target without euphemism, stating that the Federal Reserve was purchasing $1.25 trillion of agency MBS specifically "to provide support to mortgage lending and housing markets and to improve overall conditions in private credit markets." Eric Rosengren, then President of the Federal Reserve Bank of Boston, framed the selection logic in public with comparable candor in December 2012, observing that "if the purpose is to improve market functioning, or to provide focused stimulus to an interest-sensitive sector in order to stimulate aggregate demand, it may be that MBS purchases are preferable to Treasury security purchases." Both the directive and the speech named housing as the intended beneficiary. The sectoral channel was the point.
Bernanke, asked directly at the January 2009 FOMC meeting whether the program constituted credit allocation, offered a direct denial:
"I don't think that our MBS purchase program is credit allocation in the sense that this guideline was intended to address — for several reasons."
The defense was definitional: credit allocation, in Bernanke's construction, required favoring borrowers relative to normally functioning markets, and the housing market was not functioning normally. The reasoning was internally coherent. It was also operationally indistinguishable from what the dissenters were describing. The committee agreed on the mechanism and disagreed on the label. What Miran calls an improper distortion, the institution classified as the superior transmission mechanism — chose it over Treasury purchases for precisely that reason, named housing explicitly in the formal directive, and documented the logic in a March 2009 staff memo. The archive therefore confirms Miran's diagnosis: MBS purchases were sectoral credit allocation by deliberate institutional design, and the dissenters who said so in real time were analytically correct. What the archive does not resolve is the normalization question — whether the crisis-era rationale for that choice has expired sufficiently to justify clean unwinding in 2026. Internal deliberations from 2012 through 2017 show that staff consistently identified accelerated MBS sales as operationally complex and potentially disruptive, that the Committee abandoned active sales in favor of passive runoff precisely because staff judged the private-sector absorption requirement "unprecedented," and that the projected timeline for reaching a Treasury-only portfolio extended, on some staff scenarios, well past 2040. What the section establishes is not a rebuttal of the normalization prescription but its institutional context: the channel was chosen knowingly, the dissenters' diagnosis survives the archive intact, and the prescription for unwinding it inherits the same pattern of operational dependencies — market absorption capacity, distribution frictions, transition-period volatility — that constrained the other mechanisms Miran proposes. (The pattern is one the archive has surfaced before: valid diagnosis, insufficient prescription.) The institution chose the channel knowingly, for documented reasons, and against sustained internal dissent.
Miran stood before the Economic Club of Miami as a sitting Governor who helped design the proposal and brought it inside the institution. The proximity that gave his argument its credibility is the same proximity that makes the archive's response inescapable.
The investigation confirmed that the claims are testable. They have been tested. The discount window carries a stigma that the institution's own sitting leadership has characterized as structural — not administrative, not addressable through streamlined procedures or pre-positioning collateral. The LCR buffers Miran characterizes as excess regulatory demand are the insurance the institution built precisely because the window could not be relied upon; reducing one without resolving the other removes the compensation without addressing the failure it was designed to cover. The balance sheet target of one to two trillion in reduction has a direct precedent: the 2017–2019 runoff, which ended in repo market stress before reserves reached the floor the institution believed it had mapped. And the MBS channel Miran frames as an unwanted distortion was selected knowingly, against sustained internal dissent, because the alternatives served the transmission mechanism less reliably.
Each mechanism in the blueprint was not overlooked by the institution that built the current architecture. Each was examined internally; the costs, operational risks, and structural constraints were documented; and the institution chose alternatives it judged more robust under the conditions it faced. The Federal Reserve did not arrive at its current balance sheet posture by accident, by regulatory capture, or by institutional inertia. It arrived there through a series of constrained choices whose rationale the archive documents in detail — and whose reversal would require not merely administrative reform but the acceptance of operational risks and volatility costs the institution has consistently judged unacceptable.
Miran's own framing is more conditional than his critics allow. He disclaims specific advocacy, invokes cost-benefit language, and describes his options as contingent on whether reserve demand can in fact be lowered without destabilizing the system. He also argues that renewed balance-sheet reduction, if undertaken, could imply a lower federal funds rate path than otherwise — because balance-sheet shrinkage is itself contractionary. That point is analytically separate from the operational question this post has examined. If the architecture were feasible, the rate-path implication would matter. But the archive's force lies earlier in the chain: the mechanisms on which the architecture depends were examined before, found wanting in practice or in design, and compensated for by the framework now in place.
That conditionality is precisely where the archive bears most directly. The question the documentary record poses is not whether the institution was wrong to have made cost-benefit judgments — every institutional choice under uncertainty involves such judgments — but whether the evidence that drove those judgments has changed enough to warrant reversing them. On the four mechanisms the proposal names, the archive records the reasoning behind each institutional decision in unusual detail. That reasoning is the guide.
The FOMC Insight Engine provides semantic search across 90 years of Federal Reserve documents. Every claim in this article can be verified.
Explore the Archive →