The Fed Pauses RMP Buying, Testing Arthur Hayes’s ‘Stealth QE’ Thesis

As of August 14, 2026, the Federal Reserve has paused new reserve management purchases for the current operating period. The New York Fed’s current purchase schedule calls for no RMP from August 14 through September 14, after three consecutive periods of approximately $10 billion and a substantially faster pace earlier in 2026.

That pause does not settle the argument Arthur Hayes started in December 2025, but it changes the evidence. In Hayes’s December 19 essay, he portrayed RMP as an effectively unlimited successor to quantitative easing that could finance government borrowing, inflate asset prices and ultimately favor Bitcoin. The Fed is still managing a larger supply of reserves, yet the declining purchase pace—and now a zero-RMP month—shows that the program has not followed a simple, continuously accelerating money-printing path.

What reserve management purchases actually do

RMP changes the composition of private-sector balance sheets and increases reserves in the banking system. When the Fed buys a Treasury security, the seller gives up that security, while the banking system receives additional reserve balances. The Fed records both the acquired asset and the corresponding reserve liability on its own balance sheet.

That accounting resemblance to QE is real. Both operations involve central-bank purchases of securities, both can enlarge the Fed’s balance sheet, and neither requires the Fed to collect an equivalent amount of taxes or deposits before settlement. Calling the process “money printing” captures the creation of central-bank liabilities, but it does not by itself establish how much new spending, bank lending or risk-taking will follow.

The transmission depends heavily on what the Fed purchases. Treasury bills are short-dated instruments that already behave much like cash for many institutional investors. Exchanging them for reserves or deposits removes far less interest-rate risk from the market than buying ten-year notes or mortgage-backed securities, so the portfolio effect central to conventional QE is weaker.

Why the Fed says RMP is not QE

The official distinction rests on purpose, maturity and the intended market channel. In a January 16 speech, Fed Vice Chair Philip Jefferson’s account said RMP began in December 2025 after reserves had fallen to an “ample” level. He described purchases of Treasury bills and other securities with no more than three years remaining to maturity as a way to preserve control of short-term rates, while defining QE as stimulus intended to lower longer-term yields when the policy rate is constrained.

The operating problem is that reserves move when other Fed liabilities move. Tax payments can raise the Treasury General Account and drain bank reserves; currency demand and growth in the financial system can also alter the amount needed for smooth rate control. Under an ample-reserves framework, the Fed supplies enough reserves that modest daily fluctuations do not push the federal funds rate away from its target range.

This explanation supports a meaningful distinction, but it should not be stretched too far. An operation can be designed for rate control and still affect liquidity at the margin. Conversely, a larger balance sheet does not automatically deliver the broad easing associated with purchases deliberately concentrated in long-duration bonds. The relevant question is not whether RMP and QE share any accounting features—they do—but whether their scale and asset mix materially loosen broader financial conditions.

Where Hayes’s argument is strongest—and where it jumps ahead

Hayes’s strongest point is that labels should not substitute for balance-sheet analysis. RMP creates reserves through asset purchases, and its size is not governed by a single preannounced lifetime cap. If the structural demand for Fed liabilities rises, purchases can return or expand without being presented as an emergency stimulus program.

His broader thesis requires several additional links, however. Investors that sell bills must redeploy the proceeds into riskier or longer-duration assets; financing conditions must ease; Treasury borrowing must translate into additional spending; and the resulting liquidity must reach assets such as Bitcoin. Each step is possible, but none follows mechanically from the initial reserve creation.

The maturity difference is especially important. Conventional QE removes substantial duration risk from investors and is expressly intended to reduce longer-term borrowing costs. Bill purchases mainly exchange one short-duration government-linked asset for another. Hayes argued that money-market funds could redirect cash into new bills or repo lending and thereby support other Treasury buyers, but that indirect route depends on relative yields, counterparty demand and the willingness of investors to expand positions.

Nor does RMP alone prove that the Fed is directly underwriting any particular fiscal program. The purchases occur through market operations, while Treasury issuance, congressional spending decisions and the Fed’s implementation framework remain separate institutional processes. Their effects can interact, but treating interaction as a single coordinated policy goes beyond the demonstrated evidence.

What the 2026 slowdown tells investors

The purchase path provides the clearest test of the original “unlimited” framing. The New York Fed initially planned approximately $40 billion of RMP for each of four operating periods beginning in December 2025. It then reduced the amount to about $25 billion for the period ending May 13, followed by roughly $10 billion in each of the next three periods and zero for the period beginning August 14.

This sequence is consistent with a front-loaded reserve operation responding to seasonal drains, not an uninterrupted QE campaign. It does not mean RMP has been abolished: future purchases can resume if reserve demand, Treasury-account movements or money-market pressures warrant them. “Paused” is therefore more accurate than “ended.”

Investors evaluating whether RMP is becoming macroeconomic stimulus should watch more than the headline balance-sheet total. The most informative signals are the monthly amount of new RMP, the maturity of securities acquired, pressure in repo and federal-funds markets, and evidence that longer-term yields or credit conditions are moving because duration is being removed from private portfolios.

For crypto markets, RMP is best treated as one liquidity input rather than a self-contained price forecast. Hayes identified a plausible route from reserve creation to greater demand for financial assets, but the Fed’s 2026 taper demonstrates that the first link can vary sharply from month to month. The updated record supports his warning to examine central-bank plumbing, while weakening the stronger claim that RMP had already become a permanently accelerating form of stealth QE.

Also read:

• Gary Stevenson Pauses Weekly YouTube Videos After Health Concerns