The United States is running two money experiments at once. Chair Kevin Warsh is telling markets that 3.7% PCE inflation is not 2%, and that the Federal Reserve has work to do if the trend does not turn. At the same time, the Treasury is meeting a widening deficit by leaning on bills — securities that mature inside a year, carry almost no haircut in the repo market, and sit in money-market funds as if they were a close cousin of bank reserves.
Bloomberg macro strategist Simon White put a sharp label on the second experiment this week: the Treasury is beginning to look like a shadow central bank. The phrase is provocative. The plumbing is not. Bills as a share of marketable Treasury debt have been running in the low-20% area, above the unofficial 20% comfort zone Treasury officials have cited for years — White’s figures were 22.7%, or 24.1% if Fed-held bills are excluded — with the share expected to rise if coupon auction sizes stay frozen and the deficit does not. Reuters, in July, had bills at about 22% of outstanding marketable debt. Wells Fargo has sketched a path toward the mid-20s if coupons stay unchanged. That is not a rounding error. It is a duration choice that changes who feels the overnight rate.
For readers of a Canadian mining book, the question is not whether Washington’s debt managers are clever. It is whether a state that funds itself with money-like paper can still deliver a real policy rate high enough to break inflation — and what happens to gold, silver, copper and the companies that dig them if it cannot.
This is analysis, not a recommendation to buy or sell any metal or equity. Mining and bullion remain volatile. Investors can lose money.
What “Shadow Central Bank” Actually Means
Paul McCulley coined “shadow bank” in 2007 for lenders that created credit outside the regulated deposit system. White’s twist is institutional, not metaphorical. A central bank issues the instrument used for final settlement — Federal Reserve notes and reserve balances. Everything beneath that in Perry Mehrling’s hierarchy is a claim on the layer above: commercial-bank deposits, then repo, money-market fund shares, asset-backed commercial paper, foreign-exchange swaps.
Treasury bills were never supposed to live at the top of that stack. They were government IOUs with a short clock. In practice they have been migrating upward. Money funds treat them as cash. Dealers finance them in repo at a haircut that, for bills, is often zero. Collateral can be re-pledged. Academic work on the 2015–2021 period estimated that Treasury collateral in the dealer matched book was rehypothecated several times over relative to dealers’ own holdings. Zero haircut plus re-use is the definition of a money-like claim. Each turn does not subtract value the way a haircut on a long bond does. The paper multiplies.
From the Treasury’s chair, the preference is rational. Bills tap the enormous cash-management complex. They usually clear cheaper than 10-year or 30-year paper. They pull balances out of higher-velocity bank deposits. They let the government postpone the political cost of terming out debt when term premia are hostile. Recent buyback design — doubling liquidity-support operations in the 10- to 30-year sector from September, funded in the market’s base case by more short paper or by drawing the Treasury General Account — is the same impulse in another costume: suppress the long end, issue or reshuffle the front end.
From the Fed’s chair, and from anyone who owns a duration-sensitive asset, the preference is a constraint. If a growing share of the sovereign balance sheet re-prices every few weeks off the administered overnight rate, a hike is not only a monetary event. It is an immediate fiscal event. The interest bill on the public debt is already a trillion-dollar-class line item. Bills transmit that bill faster than a 30-year bond that was issued three years ago. Raise the funds rate and the Treasury’s rollover cost jumps now. That is fiscal tightening the administration did not legislate. It is also political pressure on the central bank not to legislate it with the only tool Warsh says he still treats as primary: short-term rates.
The Inflation Channel Is Not a Slogan
White’s historical sketch is the claim that jumps in bills as a share of debt have tended to precede broad inflation pulses, with the series pushed forward to line up with CPI. Correlation is not a printing press. The causal stories that survive scrutiny are narrower and more useful.
First, liquidity composition. When the government issues an instrument that money funds and dealers treat as a reserve substitute, it is expanding the stock of near-money even if the Fed’s balance sheet is not growing that week. Near-money finances asset purchases. Asset purchases raise wealth. Wealth raises spending with a lag. That is not 1970s wage-price spiral mechanics. It is a financial-conditions channel that can keep services inflation sticky after goods inflation cools.
Second, policy paralysis. Warsh’s Jackson Hole standard was confidence that underlying inflation is moving toward 2% at sufficient speed. PCE is 3.7% over twelve months and 4.1% over six on the figures he cited. Hike odds for the September 16 meeting have lived this week in the mid-60s to around 70%. That is the market taking him literally. The shadow-Treasury problem is what happens after one hike. If bills keep rising as a share of debt, each additional quarter-point becomes a larger automatic squeeze on the budget. The incentive to stop at one “credibility hike” and then pause is obvious. A pause that leaves 3.7% in the rear-view is how inflation expectations stop falling. Metals traders have a name for that regime. They call it poorly anchored.
Third, the money-fund and repo loop. Extra bill supply can cheapen bills relative to the overnight reverse-repo facility and to bank reserves. Funds rotate. When the RRP buffer is already thin — it was a few billion dollars at the end of August after a multi-trillion peak years earlier — the next rotation lands on reserves and private repo. Funding flare-ups are not inflation in the CPI. They are the financial-stability half of the Fed’s job colliding with the inflation half. A central bank that is putting out a plumbing fire is not hiking into 4% six-month PCE. That collision is inflationary in effect even when it is defensive in intent.
Fourth, the wealth-and-capital-cost loop White flags directly. More liquidity under risk assets represses the cost of capital, bids up claims on future cash flows, and feeds back into rents, financial services, and the PCE basket Warsh is watching. The Fed then faces a market that is not tight even when the funds rate looks restrictive on a Taylor-rule slide.
None of this requires a conspiracy. It requires a deficit, a preference for bills, and a central bank whose main tool is the same rate that now prices a larger slice of the government’s own liabilities.
Why the Metals Tape Can Fall First and Rise Later
Canadian mining investors have just watched the first half of that sentence.
Gold gave back a violent piece of August after Jackson Hole and opened September with another down-leg as the 10-year yield pushed toward the high-4.70s. Silver, which had used the mid-$66s as a shelf, broke through $66 on September 1 and printed the mid-$64s to high-$63s on subsequent screens — still roughly 45% below the winter peak near $115–$122, still 50–60% above year-ago levels near $40. The PHLX gold-silver equity index was slashed on September 1 and tried to bounce into September 2. Copper had made a COMEX record near $6.78 a pound in late August, then had to digest the same dollar-and-yield shock plus the odd two-market structure of U.S. tariff stockpiles versus tight London stocks.
That is the textbook metals response to a credible hike threat. Bullion pays no coupon. Real yields are the opportunity cost. A 70% September hike probability is a reason to sell the metal this week. It is not a reason to retire the fiscal thesis.
The second half of the sentence is the one White’s framework forces onto a mining desk. If the Treasury’s bill share keeps climbing, the Fed’s ability to deliver a path of rates that actually restores 2% without a fiscal accident shrinks. Markets will eventually treat that shrink as a change in the monetary regime, not as one more data print. Regime changes are what re-rate gold as a reserve asset rather than as a rate-sensitive commodity. They are what keep official-sector buying rational even when real yields are not falling in a straight line. They are what make silver’s industrial deficit and gold’s monetary bid rhyme again after a year of fighting each other.
Timing is the entire risk. A shadow-Treasury inflation impulse can take quarters to show up in core services. A Warsh hike can show up in COMEX before lunch. Positioning for the second without surviving the first is how January 2026’s gold crash gets a sequel.
Gold: Insurance Against a Rate That Cannot Be Set
White’s closing line is the gold thesis in bureaucratic language. Finding a policy rate that satisfies the inflation mandate, keeps funding markets orderly, and still lets the government borrow on a scale that is neither inflationary nor destabilizing may not be possible. Gold is the asset that exists for that sentence.
It does not require the Fed to fail next month. It requires the market to assign a rising probability that the failure is structural. Bill-heavy financing is one of the cleanest structural markers available, because it is observable in refunding statements and in the Monthly Statement of the Public Debt. Watch the bill share, the coupon-auction guidance, the buyback calendar, and the TGA. If coupons stay frozen and bills absorb the deficit, the gold argument strengthens even if the next PCE print is friendly. If Treasury terms out debt into 10s and 30s in size, the argument weakens, because duration risk returns to the private sector and the Fed gets a clearer shot at hiking without instant fiscal blowback.
Canadian gold producers — Agnico, Kinross, Wheaton as a streamer, the Ontario and Quebec developers — live in both clocks. The near-term clock is the September FOMC and the CAD, which often firms when U.S. yields jump and risk comes off. The long clock is whether a world that doubts the Fed-Treasury separation pays a higher real price for ounces in safe jurisdictions. Those clocks can point opposite directions in the same quarter. A miner’s board that hedges the first clock and ignores the second is running a different company from one that does the reverse.
Silver: The Metal That Gets Punished for Being Both Things
Silver remains the awkward child of this setup. It is monetary enough to fall when hike odds go to 70%. It is industrial enough that a fifth or sixth consecutive physical deficit does not vanish because the 10-year yield added 15 basis points. World Silver Survey arithmetic still described a market that mines roughly 844 million ounces and runs a tens-of-millions-of-ounces shortfall after a cooler solar year. That stock cannot reprice in a week. Paper can.
A bill-driven, inflation-tolerant regime is silver-friendly on a two-year view if industrial offtake holds and investment demand returns. It is silver-hostile on a two-week view if money funds are buying bills instead of metal ETFs and if CTAs are selling strength in real yields. Canadian silver names — the Pan American, First Majestic, Hecla-adjacent stories, the TSX Venture developers — will trade the two-week view until they are forced not to. Anyone calling a $64 print a structural bargain because “the Treasury is a shadow Fed” is collapsing those horizons. The collapse is how accounts get margin-called in September and miss a 2027 rerating.
Copper: Different Metal, Same Constraint
Copper does not care about Mehrling’s hierarchy in the way gold does. It cares about grids, data centers, EVs, Chinese fabrication, Chilean pits, and, this year, a U.S. tariff threat that vacuumed metal into COMEX warehouses at record tonnage while London felt tight. A shadow-Treasury inflation problem matters to copper in two indirect ways.
If the Fed cannot hike enough, nominal demand and infrastructure spending stay supported. That is bullish for copper demand and for the mid-tier Canadian producers and developers that sell into that demand — Hudbay, Capstone, First Quantum, Teck’s copper slice, the Arizona-and-Quebec junior file.
If the Fed hikes once for credibility and then the bill machine plus a funding scare forces a pivot, the path is a growth wobble first and a liquidity wave second. Copper hates the wobble. It likes the wave only after the wobble has already marked down the equities. That sequence is why copper mining stocks can look cheap at $6 copper and still fall 20% on a U.S. recession scare that was caused by the attempt to restore 2%.
Copper M&A, which already produced Hudbay’s Arizona Sonoran purchase and a Mitsubishi cheque into Copper World, does not pause because Simon White wrote a hierarchy chart. It pauses if the cost of capital spikes and the dollar stays bid. A Treasury that has made the whole state book overnight-rate sensitive is, in that sense, a cost-of-capital event for every PEA on SEDAR.
The Canadian Overlay: Currency, Costs, and Who Gets Paid
A U.S. fiscal-monetary fusion is not automatically a gift to Toronto-listed miners.
The Canadian dollar still trades as a risk-and-commodity currency. A Warsh hike that lifts U.S. real yields can cap CAD and help Canadian-dollar cost bases in the short run. A later regime in which global investors flee the dollar’s fiscal story can lift gold in USD and lift CAD with commodities, squeezing local-cost margins even as the metal looks higher on a Bloomberg screen. Margin is ounces times price minus CAD costs. The shadow-Fed thesis changes both sides of that identity at different speeds.
Jurisdiction premia also move. If U.S. policy looks less separable from the Treasury’s funding needs, “allied, rule-of-law, permit-able” pounds and ounces in Quebec, Ontario, British Columbia, and the Yukon become easier to defend in a board memo. That is already visible in the 2026 copper deal tape. It does not make every TSX Venture ticket a takeover. It does change which files get the first call.
Royalty and streaming companies sit in a cleaner seat than operators if the thesis is “the price of money is politically capped.” They clip the metal. They do not run the mill when energy and labor inflate because bills are money. Operators need the inflation to land in the metal faster than it lands in diesel and wages. That race is older than T-bill shares. The new element is only that the Fed may be less free to stop the race.
What Would Falsify the Metals Reading
A serious desk writes the kill criteria down.
Treasury raises coupon auction sizes in a meaningful way and lets bills fall back toward or below 20% of marketable debt. Duration returns to the private sector. The Fed’s rate tool is less fiscally radioactive.
Core PCE and the six-month trend actually break toward 2% without a funding accident. Warsh gets the inflation mandate without proving White’s impossibility claim.
Money-market plumbing absorbs another trillion of bills without reserve scarcity or a spike in repo. The “near-money” inflation channel stays a theory.
Official gold buying fades and Western ETFs keep leaking through the hike. The monetary bid was a 2022–2025 story, not a 2026–2028 story.
Until one of those prints, the working hypothesis for a mining reader is dual: respect the hike tape this month; respect the financing regime for the cycle.
How to Sit in the Book Without Pretending to Know September 16
Separate the sleeves. A trading sleeve can fade metals into 70% hike odds and buy a failed break of obvious levels. A reserve sleeve sized as insurance against fiscal-monetary fusion does not need a daily view on FedWatch. Mixing them is how people sell gold in January at the high and sell it again in September at the wrong low.
Do not let a macro essay pick a junior. White’s hierarchy does not grade a drill hole in the Golden Triangle or a PEA in Arizona. It changes the discount rate and the terminal inflation assumption that sit underneath those models. Update the models. Do not outsource project risk to a Bloomberg column.
Watch the documents that actually move the regime: the quarterly refunding, TBAC minutes, bill share of marketable debt, RRP balances, TGA, and whether Warsh’s “no regular forward guidance” Fed still hikes when the interest bill screams. Those are primary sources. The label “shadow Fed” is a headline on top of them.
Conclusion
The Treasury is not the Federal Reserve. It does not set the funds rate and it does not issue legal tender. It is, however, issuing more of an instrument that the private system treats as a stand-in for central-bank money, and it is doing so while inflation is still a full point and a half above target. That combination makes the overnight rate a fiscal switch as well as a monetary one. A switch that expensive is harder to flip — and harder to flip repeatedly.
For inflation, the risk is not a 1970s reprint next Tuesday. It is a ceiling on how tight policy can become before the government’s own rollover schedule objects. For metals, the near-term risk is the opposite: Warsh flips the switch once, yields jump, gold and silver take the punch they took this week, copper catches a dollar shock. The cycle risk is that the switch cannot stay flipped, 2% remains a slogan, and ounces in the ground in this country become the cleaner claim on a monetary system that has fused its fiscal and central-bank balance sheets at the front end.
Canadian mining investors do not need to settle the academic argument about hierarchies of money. They need to know which clock they are underwriting. This month’s clock is September 16. The longer clock is the bill share of U.S. debt. Both are now part of the same price.
Important information
This article is for informational and educational purposes only. It does not constitute investment advice, tax advice, legal advice, or a recommendation to buy, sell, or hold any metal, mining equity, ETF, bond, bill, or derivative. Arguments attributed to third-party strategists, including Simon White of Bloomberg, are presented for discussion and may be incomplete or incorrect. Treasury bill shares, Fed-funds probabilities, inflation prints, and metals prices cited here reflect public reporting as of early September 2026 and will change. Forward-looking statements about inflation, Federal Reserve policy, fiscal dominance, and commodity prices are uncertain. Mining and exploration companies involve a high risk of loss, including possible loss of principal. Readers should verify primary filings and official data — Treasury, Federal Reserve, Statistics Canada, SEDAR+, EDGAR — and consult licensed advisers. The author and publisher accept no liability for actions taken on the basis of this article. Past performance is not indicative of future results. This communication does not consider any individual’s circumstances.