What is Section 13(3)? The Fed power crypto cannot use

When the Fed chair was asked whether the central bank would rescue crypto in a crisis, he avoided one topic with lawyerly care: Section 13(3), the emergency lending power behind every modern bailout. Here is what it is, how 2010 rewrote it, and why a stablecoin rescue through it is closer to illegal than unlikely.

Every modern financial rescue Americans can name ran through the same fourteen words of statute. Bear Stearns, AIG, the alphabet of 2008 facilities, the pandemic programs of 2020: all of them drew their legal authority from the third paragraph of Section 13 of the Federal Reserve Act, which permits the Fed, in unusual and exigent circumstances, to lend beyond the banking system it normally serves. So when Representative Brad Sherman asked the new Fed chair on July 14 whether crypto would get the treatment money market funds got in 2008, and Kevin Warsh answered that the Fed does not want to be in the bailout business while conspicuously declining to discuss his emergency authority, the exchange pointed directly at this provision. For readers tracking the exchange, crypto.news has also covered the testimony this statute sits under. Most of the crypto market has never read it. That is worth fixing, because what the statute actually says, especially after Congress rewrote it in 2010, changes the question from whether the Fed would rescue a stablecoin issuer to whether it lawfully could. The answer is narrower than almost anyone trading on the assumption of a backstop believes.

Where the power came from

Section 13(3) is a Depression artifact, and its origin explains its shape.

The Federal Reserve of 1913 was built to lend to banks against short-term commercial paper, a deliberately narrow design. The Great Depression broke the design’s assumptions: thousands of banks failed, surviving banks hoarded, and creditworthy businesses could not borrow at any price. Congress responded with the Emergency Relief and Construction Act of July 1932, adding a third paragraph to Section 13 that let the Reserve Banks lend to individuals, partnerships, and corporations, anyone, in effect, when circumstances were unusual and exigent, the borrower could post satisfactory collateral, and at least five members of the Federal Reserve Board approved. Historians of the provision note that its framers meant it to reach the real economy, not merely a weakened financial sector: it was a tool for lending to merchants when the banking system had seized.

Then it went to sleep. The authority sat essentially unused for three-quarters of a century, a loaded but forgotten instrument, until 2008.

What 2008 did with it

The financial crisis turned Section 13(3) from a footnote into the operating system of the rescue.

The Fed invoked it in two distinct ways, and the distinction is the entire modern debate. The first was broad-based: facilities open to whole classes of borrowers, created to revive whole markets. Programs for primary dealers, for commercial paper, for asset-backed securities, six facilities designed, five used, all justified as providing liquidity to the financial system rather than saving anyone in particular. The second was tailored: special assistance built for exactly one counterparty at a time. A $13 billion direct loan to Bear Stearns and roughly $30 billion more to grease its sale to JPMorgan. The AIG rescue. Support arrangements for Citigroup and Bank of America. Four firms the Fed judged too big to fail, each receiving a bespoke intervention under the same fourteen words written for Depression-era merchants.

The tailored rescues worked, in the narrow sense that the firms did not collapse, and they poisoned the politics of the authority, in the broad sense that Congress concluded a central bank should never again design a private rescue for a chosen firm. That conclusion became law.

How Dodd-Frank rewired it

The 2010 Dodd-Frank Act did not repeal Section 13(3). It did something more interesting: it kept the power and removed the part crypto is implicitly counting on.

The amendments, implemented in a final Fed rule in 2015, impose five binding constraints. Emergency lending must be through a program or facility with broad-based eligibility, meaning open to a class of borrowers, designed to supply liquidity to the financial system, and explicitly not for the purpose of aiding a single failing financial company. Borrowers must be solvent; the Fed is required to maintain procedures prohibiting credit to insolvent firms. Collateral must be sufficient to protect taxpayers from losses. The Treasury secretary must approve any program before it launches. And the whole exercise runs under mandatory disclosure with a lag plus Government Accountability Office audit.

Read those constraints against the 2008 record and the intent is unmistakable: the broad facilities would have been legal under the new rules, and Bear Stearns, AIG, Citigroup, and Bank of America would not. Congress banned the bespoke bailout while preserving the market-wide fire hose. The 2020 pandemic response proved the surviving architecture works as designed: the Fed reopened its broad facilities and built new ones, corporate credit, municipal liquidity, Main Street lending, all broad-based, all Treasury-approved, several capitalized with Treasury equity that the Fed leveraged, and none of them a rescue of any single named firm.

Now apply it to crypto

Walk a stablecoin crisis through the modern statute and the market’s implicit assumptions start failing the text.

Scenario one: a major issuer breaks. Its coin depegs, redemptions surge, and its reserves, wherever they sit, cannot be liquidated fast enough. Crypto.news has explained what a run on an issuer looks like in stablecoin markets. Could the Fed lend to the issuer to bridge the run? Under post-2010 law, almost certainly not. A loan to one named issuer is precisely the single-firm assistance Dodd-Frank prohibits; a facility gerrymandered to reach only that issuer would be the same thing in costume, which the 2015 rule anticipates. And an issuer whose liabilities exceed the realizable value of its assets in the relevant window has a solvency problem, which triggers the categorical bar. The legal analysis is not close. The tool the market imagines, the Fed catching a falling Tether or Circle the way it caught AIG, was welded shut fifteen years ago.

Scenario two: the sector runs, not one firm. A generalized stablecoin panic forces mass liquidation of reserve assets, Treasury bills and repo, at fire-sale speed, and the stress starts transmitting into the funding markets banks and money funds share, which is exactly the channel the New York Fed’s staff research has flagged. Here a lawful path exists: a broad-based facility lending against high-quality reserve assets to a defined class of participants, justified as protecting the Treasury and money markets rather than any issuer. It would need Treasury sign-off, five board votes, taxpayer-protective collateral, and eventual disclosure, and it would look less like saving crypto than like the Fed defending the government securities market with crypto as an incidental beneficiary.

Which is the precise shape of Warsh’s July 14 testimony. His full stop, no bailouts, maps onto what the statute already forbids: firm-specific rescue. His hedge, mitigating extraordinary risks, maps onto what the statute still permits: broad liquidity defense of the system. The chair avoided discussing 13(3) not because the answer is embarrassing but because the answer is the law, and stating it plainly, we legally cannot save your issuer, and might flood the market it drowns in, is not a sentence any central banker volunteers.

One more concreteness is worth adding before leaving the crypto scenarios, because the abstract phrase broad-based facility hides real design choices that would decide who actually benefits. A lawful stablecoin-crisis facility would have to define its borrower class, and every plausible definition changes the politics. A facility lending to banks against Treasury collateral, the 2023 template, helps issuers only indirectly, by keeping the bill market orderly while they liquidate. A facility lending to registered stablecoin issuers as a class against their reserve assets would be legally defensible under the broad-based test once the GENIUS regime defines who a permitted issuer is, and it would instantly raise the question Congress fought over in 2008: why this industry’s liquidity and not another’s. A facility reaching exchanges or custodians would strain the financial-system purpose language and almost certainly fail the Treasury-approval gate. The unfinished GENIUS rulebook matters here too, in an underappreciated way: a facility for permitted payment stablecoin issuers is only definable once the licensing rules say who they are. The missed July deadline did not just delay compliance paperwork. It delayed the existence of the borrower class any lawful crypto facility would need, which means that today, in a crisis, even the legal path would begin with regulators improvising definitions, the exact condition emergency lending law was rewritten to prevent.

The rescue that confused everyone

One episode makes the market chronically overestimate the crypto safety net, and it deserves to be filed correctly: March 2023, when USDC broke and was made whole.

That was not Section 13(3), and it was not the Fed acting as lender of last resort to crypto. Circle held $3.3 billion of USDC reserves as deposits at Silicon Valley Bank; the bank failed; the coin fell to roughly 87 cents. What restored it was the FDIC’s systemic risk exception, the different tool that actually rescued crypto once, a separate authority at a separate agency under separate law, which allowed regulators to guarantee all SVB depositors, uninsured ones included, to stop a regional banking contagion. Circle was a depositor, so Circle was caught in the net, so the peg recovered. The Fed’s contribution that weekend was a new lending facility for banks, broad-based, exactly as Dodd-Frank prescribes.

The correct lesson is double-edged. Crypto’s one historical rescue was an accident, a spillover from the traditional system saving itself, and the specific channel it flowed through, uninsured issuer deposits at a bank, is precisely the exposure the post-2023 reserve reforms and the GENIUS Act’s rules are designed to shrink. The accidental-bailout pathway is narrowing by design. What remains, on the Fed side, is only the broad facility, with its political gate at Treasury and its solvency screen at the door.

The money market fund precedent, examined

The exchange that produced Warsh’s testimony began with a specific historical reference, Sherman asking whether crypto would get what money market funds got in 2008, and the comparison rewards a closer look, because it is simultaneously the strongest argument for crypto’s eventual rescue and the strongest argument against it.

What money market funds got in 2008 was not, strictly, a Section 13(3) loan to a failing fund. When the Reserve Primary Fund broke the buck after Lehman’s collapse and a run began across the industry, the response came in two parts. Treasury created a temporary guarantee program for money fund shares, backed by its own Exchange Stabilization Fund, effectively insurance conjured overnight for an uninsured product. The Fed, for its part, built broad-based 13(3) facilities that lent against the assets funds were dumping, restoring the markets the funds needed to meet redemptions. Firm-specific rescue never happened; system-wide liquidity and an improvised guarantee did, and together they stopped the run within weeks.

The parallel to a future stablecoin crisis is close enough to be uncomfortable. A stablecoin is functionally a bearer money market share: a claim on a pool of short-dated assets, promising par, redeemable on demand, held by users who treat it as cash. A sector-wide stablecoin run would look like September 2008 in miniature, mass redemption, fire sales of bills and repo, contagion through whatever the coins collateralize. And the toolkit that worked then maps onto what remains legal now: the Fed could lawfully build a broad facility against reserve assets, exactly as it did for the funds’ assets, and Treasury retains its own instruments outside the Fed’s statute entirely. Anyone reasoning from 2008 concludes that the system, pressed hard enough, finds a way, and that conclusion is not naive. It is the historical base rate.

But the aftermath of 2008 is the other half of the precedent, and it points the opposite way. The money fund rescue was followed by fifteen years of regulatory effort to ensure it never recurred: floating net asset values for institutional funds, liquidity fees, gates, reform fights in 2014 and again in 2023, all animated by the conviction that an uninsured product which received an improvised guarantee once must be restructured so it never needs one again. The rescue bought the industry survival and cost it the presumption of independence. Stablecoins are receiving the sequel in advance: the GENIUS Act’s full-reserve and holder-priority rules are the money fund reforms applied before the crisis instead of after, a legislature attempting to pre-position the orderly-failure machinery so the improvised-guarantee moment never arrives.

Which resolves the Sherman question more precisely than either a yes or a no. Would crypto get what money market funds got? The firm rescues, never, those are barred. The broad liquidity, plausibly, that door remains open by design. The improvised guarantee, only at the point where a stablecoin run visibly threatens the Treasury market itself, and the entire current regulatory project is an attempt to make sure the question is never asked, by making failure survivable before it happens. The 2008 precedent is real, and it comes with its own warning label: the products that used it spent the next decade paying it back.

Why the narrowness is the point

It is tempting to read all this as crypto being uniquely disfavored. The truth is closer to the opposite: crypto is being handed, in advance and in writing, the exact deal the rest of finance learned the hard way.

The GENIUS Act is the resolution regime that makes no-bailout credible. Its full-reserve requirement and its rule paying stablecoin holders ahead of other creditors in an insolvency are the components of orderly failure, the thing a system needs so that firms can die without rescues. A resolution regime and a constrained lender of last resort are complements: the first makes the second credible. The unfinished state of the GENIUS rulebook, all six agencies having missed the July 18 rulemaking deadline, is therefore not a side story. Crypto.news has also covered the unfinished rulebook behind the doctrine. Until redemption mechanics and supervisory triggers are final, an issuer failure would be improvised, and improvisation is historically where no-bailout doctrines go to die. The statute bars the tailored rescue; only a working resolution process bars the pressure for one.

For anyone holding or building in the sector, the practical summary is short. There is no lawful mechanism for the Fed to rescue your issuer, your exchange, or your custodian as such. There is a lawful mechanism for the Fed to flood the markets your issuer’s reserves live in, if a failure ever threatens those markets, and using it requires the Treasury secretary’s signature and a solvent counterparty class. Everything else, reserve quality, segregation, attestation, legal priority, is the actual safety net, and it is private. Section 13(3) is the most famous emergency power in finance, and the most important fact about it for crypto is fourteen years old: Congress already decided who it cannot save.

Frequently asked questions

Disclaimer: This article is for information and educational purposes only and does not constitute financial, investment, or legal advice. It summarizes statutes and regulatory practice that are subject to interpretation and change, and no description here should be relied on as a prediction of official action. Always do your own research. Information is accurate as of July 20, 2026.