The inflation report arrived this morning wearing a respectable suit.
The headline Producer Price Index rose 0.4 percent in August, and 5.4 percent over the year. Final-demand services rose 0.1 percent. The first reading was familiar: prices were rising, but by the close the market was bracing for a hotter CPI print and marking up September rate-hike odds after the PPI surprise landed on the upside.[1]
Then look underneath the headline.
Regatta’s morning file, using the older production-pipeline lens, recorded processed goods for intermediate demand up 11.5 percent over the year, unprocessed goods for intermediate demand up 12.8 percent, and diesel up 24.1 percent on the file’s stated monthly basis. The same file recorded 206,000 initial jobless claims for the Labor Day week. Those intermediate-goods, diesel, and claims figures must be verified against the underlying Bureau of Labor Statistics and Department of Labor tables before publication; the precise periods and line items matter.[2]
Gold traded near $4,390 and silver near $64–65 by the September 10 close; after the hot-PPI selloff, gold closed near $4,392 and silver near $64.58, with silver down about 5.9 percent on the day.[2]
The result is not necessarily a contradiction. It is a measurement problem.
One inflation story is being told by the modern final-demand index. Another is being told by the older pipeline of raw materials, processed inputs, and industrial costs. Both can be true because they are not measuring the same thing.
That distinction matters. The PPI published today is not the PPI that sat beneath the gold market’s 1980 peak. The CPI used to compare today’s gold price with 1980 is not quite the CPI a 1970s household would have recognized, either.
The ounce did not change.
The yardsticks next to it did.
This Is Not a Conspiracy. It Is a Chronology.
The word “manipulation” is often used carelessly. A government agency changes a formula, and someone declares that the numbers have been falsified. That is too simple. Statistical agencies have real reasons to revise their methods: consumers substitute, products improve, industries change, and the economy moves from factories toward services.
But the opposite mistake is just as serious: assuming that a number remains conceptually unchanged because its name remains the same.
A rebuilt index may be more technically sophisticated and still answer a different question.
The modern CPI may be better at estimating a national cost-of-living trend than its predecessor. It may be less useful for answering the question a family asks at the grocery store: How much more money must leave my account this month to live as I lived last year?
The modern PPI may be better at measuring price movements across a services-heavy economy. It may be less useful for answering the question a mining company asks: Are diesel, steel, labor, and other inputs becoming more expensive than the ounce of metal we sell?
The honest complaint is not that every statistic is fake. It is that the official statistics have been rebuilt, openly and repeatedly, to serve changing analytical purposes.
The Inflation Clock Was Reset
The first major change concerns housing.
Beginning with January 1983 data, the CPI shifted its treatment of owner-occupied housing toward owners’ equivalent rent. Instead of treating a homeowner’s purchase price, mortgage interest, and house-price movements as current consumption costs, the index asked what the home would rent for in the market.[3]
That was an important conceptual decision. A house is both a place to live and an asset. The CPI chose to measure the consumption service, the shelter, rather than the investment value.
The decision was defensible. It was also transformative.
A homeowner may experience a sharp rise in property taxes, insurance, repairs, mortgage payments, or the replacement cost of a roof. Owners’ equivalent rent may not move in the same way, at the same speed, or even in the same direction. The index is not “wrong” because it does not mirror every household. It is measuring an imputed rental service rather than the household’s full cash burden.[3]
Then came the Boskin Commission in 1996. The commission argued that the CPI overstated inflation because it did not fully account for substitution, changing retail outlets, new products, and improvements in quality.[4] BLS subsequently expanded the use of hedonic quality adjustments, changed basic-item aggregation, introduced a geometric-mean formula for many components in 1999, and introduced the chained CPI in 2002.[5]
The logic is straightforward. If a computer becomes faster at the same price, the statistical model can treat part of the apparent price stability as an improvement in quality. If shoppers switch from expensive steak to cheaper chicken, a substitution-sensitive measure can record a smaller increase in the cost of maintaining consumption.
Again, the methods may be reasonable. But the number becomes less like a cash register and more like a model of economic welfare.
CPI weights also changed. BLS moved to biennial weight updates beginning in 2002 and began annual updates for the CPI-U and CPI-W in the early 2020s, including the 2023 transition to annual expenditure weights.[5] The modern index therefore reflects a newer consumption basket than the index used in 1980.
So when someone says gold is “up” or “down” relative to its 1980 inflation-adjusted value, the first question is not whether the arithmetic is correct.
The first question is: which inflation series did you use?
The Producer Index Was Rebuilt, Too
The PPI’s history is even more revealing.
Before 1978, the principal measure was the Wholesale Price Index. Its broad commodity indexes could count price changes multiple times as a product moved from crude material to intermediate processing to finished goods. The program was renamed the Producer Price Index in 1978, and BLS emphasized the Stage-of-Processing framework: crude materials, intermediate goods, and finished goods.[6]
That pipeline was useful because it followed the physical river of production.
Oil became fuel. Fuel powered machinery. Machinery and labor produced manufactured goods. Rising costs moved downstream, sometimes slowly and sometimes violently. Analysts in the 1970s could watch pressure build in the intermediate stages before it appeared in finished goods.
In 2014, BLS made another major change. The Final Demand–Intermediate Demand system replaced Stage of Processing as the primary aggregation structure. Final demand includes goods, services, construction, government purchases, and exports sold for final use. Intermediate demand covers goods, services, and construction used as inputs into further production.[7]
This was not a minor relabeling. It changed what the headline was designed to summarize.
The modern final-demand index is a broad blend of an economy in which services, trade margins, construction, government procurement, and exports matter alongside goods. A large increase in one part of the economy can be offset by weakness elsewhere. The result can be a relatively calm headline even while a particular production chain is under severe stress.
For Regatta’s analytical purposes, the intermediate processed and unprocessed goods series are the “original PPI” lens: not the official name of the index, but a useful description of the older pipeline question. The final-demand headline is the post-2014, CPI-cousin version: a broader measure designed to capture the price environment facing final purchasers.
Neither is inherently more truthful. They simply answer different questions.
That is why a final-demand reading of +0.4 percent for the month can coexist with much hotter movement in intermediate materials. The headline may be contained while the foundation is being repriced.
For a metals analyst, the crucial question is not “Is CPI contained?” It is:
Is an ounce of gold buying more diesel, labor, and steel than it did 12 months ago?
That is a pre-1980 PPI question. It belongs to the pipeline.
The Money Tape Was Spliced
The money supply has its own broken chronology.
In March 2006, the Federal Reserve discontinued publication of M3 and several of its components, including large-denomination time deposits, certain repurchase agreements, and Eurodollar deposits. The Fed said M3 provided little additional information beyond M2 relative to the cost of producing it.[8]
That explanation may be entirely sincere. But the analytical consequence remains: a broad monetary aggregate disappeared from the regular public dashboard just as nonbank finance and shadow-banking structures were becoming increasingly important to the financial system.
M3 did not vanish from history. Historical estimates remain available. But a chart comparing modern “money printing” with gold using M2 alone is not automatically comparable with an older chart that included a broader set of money-like liabilities.
Then, in 2020, M1 was redefined. Savings deposits were moved into M1 after regulatory changes made them more transaction-like, and the Federal Reserve introduced the category “other liquid deposits.” M1 consequently jumped on paper, not because every dollar of savings had suddenly become newly printed cash, but because the classification changed.[9]
This is the statistical equivalent of moving water from one reservoir label to another.
The water matters. The label matters, too.
The practical rule is simple: when someone presents a chart showing money supply and gold, ask:
- Is the series M1, M2, M3, or a reconstructed aggregate?
- Are the definitions consistent over the full period?
- Did the series change because the economy changed, or because the category changed?
- Are bank deposits being treated as money, liquidity, or credit?
Good decisions require more than a rising line labeled “money.”
The Dollar Ruler Is a 1999 Artifact
“The dollar is strong” is another sentence that sounds precise until you ask: strong against what?
The U.S. Dollar Index, or DXY, was created before the euro. Its present six-currency composition dates from the euro’s introduction in 1999. The basket assigns approximately 57.6 percent to the euro, 13.6 percent to the yen, 11.9 percent to sterling, 9.1 percent to the Canadian dollar, 4.2 percent to the Swedish krona, and 3.6 percent to the Swiss franc.[10]
There is no Chinese renminbi, Mexican peso, Korean won, Indian rupee, or Brazilian real in that basket.
DXY is therefore a ruler with a particular shape. It is useful for measuring the dollar against six currencies. It is not a complete measure of the dollar’s purchasing power against the world, and it is certainly not a measure of the dollar against gold.
A dollar can be stable or strong on DXY while losing value against gold, oil, housing, insurance, or industrial inputs. Those are not mutually exclusive observations.
Every time a report says “dollar strength,” specify the instrument:
- DXY strength means strength against the fixed six-currency basket.
- Gold strength means the dollar is buying fewer ounces.
- Trade-weighted dollar strength depends on which trading partners and weights are included.
- Household purchasing power depends on the goods and services a household actually buys.
Confusing these rulers is how a portfolio can be described as protected while its real purchasing power quietly erodes.
Even the Precious-Metals Benchmarks Modernized
Gold and silver have not been frozen in time. Their physical units are stable, but the markets through which prices are discovered have changed.
The London Silver Fix was replaced by the LBMA Silver Price on August 15, 2014. Platinum and palladium benchmarks moved to new electronic arrangements on December 1, 2014. On March 20, 2015, the London Gold Fix was replaced by the LBMA Gold Price, administered by ICE Benchmark Administration through an electronic auction process.[11]
That transition represented a move toward greater auditability and formal benchmark governance. It did not make earlier prices unusable. But it did change the market structure behind the published number.
A troy ounce remains a troy ounce: precisely 31.1034768 grams under the LBMA’s technical conversion factor, commonly rounded to 31.1035 grams.[12] The ounce cannot be subjected to a hedonic adjustment because the bar became shinier. It cannot be reweighted because households substituted into a cheaper ounce. The physical unit does not improve its quality while remaining the same weight.
That is why gold and silver can serve as useful rulers, not perfect rulers, and not short-term trading signals, but rulers that are harder to redefine by committee.
There is still a distinction between the physical metal and the market infrastructure around it.
On COMEX, registered metal has a warehouse warrant and is deliverable against futures contracts. Eligible metal meets contract specifications but does not currently have a warrant. Total stocks combine both categories.[13]
So when someone says “COMEX inventories collapsed,” ask:
Registered, eligible, or total, and did total metal actually leave the vault?
Registered metal can become eligible when a warrant is canceled. Eligible metal can become registered when a warrant is issued. The category can change without the bar leaving the building.
The Paper Market's Rap Sheet: When the Price-Setters Work Against Their Own Clients
Before anyone turns this into folklore, start with the receipts. The enforcement record is not a rumor passed around a bullion forum. It is a matter of public orders, deferred-prosecution agreements, final notices, and court-approved settlements. The institutions below were not fringe operators wandering in from the parking lot. They were core participants in precious-metals price discovery, and regulators found repeated episodes of spoofing, attempted manipulation, benchmark abuse, false statements, and supervisory failure.[14][15][16][17][18][19][20][21][22][23][24][25][26][27]
| Institution | Resolution | Conduct documented in public record | Primary sources |
|---|---|---|---|
| JPMorgan | $920.2 million (September 2020) | CFTC, DOJ, and SEC resolution covering spoofing and manipulation in gold, silver, platinum, palladium, and U.S. Treasury futures; DOJ said 15 traders across two desks were involved; the bank admitted wrongdoing and entered a three-year deferred prosecution agreement.[14][15] | CFTC 8260-20; DOJ press release |
| Deutsche Bank | $30 million (January 2018) | CFTC found manipulation, attempted manipulation, and spoofing in COMEX precious-metals futures from 2008 to 2014.[16] | CFTC 7682-18 |
| Bank of Nova Scotia (Scotiabank) | $127.4 million in combined penalties (2020) | CFTC actions covering spoofing, false statements, and supervision/compliance failures.[17][18] | CFTC 8220-20; CFTC 8221-20 |
| UBS | $15 million (January 2018) | CFTC found attempted manipulation and spoofing in precious-metals futures.[19] | CFTC 7683-18 |
| HSBC Securities (USA) Inc. | $1.6 million (January 2018) | CFTC found spoofing in precious-metals futures.[20] | CFTC 7684-18 |
| Morgan Stanley Capital Group | $1.5 million (2019) | CFTC found spoofing in precious-metals futures.[21] | CFTC 8044-19 |
| Barclays | £26 million (about $44 million at the time, May 2014) | UK FCA fined Barclays for failings around the London Gold Fixing after a trader manipulated the June 2012 gold fix to benefit the bank’s own options position.[22] | FCA final notice / press materials |
| Moore Capital Management | $25 million (2010) | CFTC found attempted manipulation of platinum and palladium futures settlement prices through “banging the close”; a related portfolio manager later paid $1 million; Reuters later reported a $48.4 million private settlement in 2013.[23][24][25] | CFTC 5815-10; Reuters |
| Private antitrust litigation | Roughly $184 million in combined settlements | Public litigation over gold and silver benchmark abuse produced large settlements across institutions, including a $60 million Deutsche Bank settlement in the Gold Futures and Options Trading Litigation and a $38 million Deutsche Bank settlement in the London Silver Fixing litigation.[26][27] | Gold litigation settlement filings; Court filing on Deutsche Bank gold settlement |
The point is not merely that bad acts occurred. The point is who committed them. These were not anonymous outsiders trying to mug the market in an alley. These were the same institutions sitting near the center of COMEX and LBMA price discovery: the firms whose traders helped shape the daily benchmark, whose desks held client orders, whose vault and custody arms held metal, and whose research departments advised the same families and institutions trading that metal. The enforcement record shows traders at those firms manipulating the prices of the very instruments their clients were trading — placing spoofed orders with intent to cancel, or leaning on a fix to benefit the bank’s own book.[14][16][19][20][22] That is the modern structural conflict in plain English: the price-setter, the order-taker, and the advice-giver are the same firm, and when the firm’s own book is on the other side, the client is the counterparty. This is why Regatta Financial operates fee-only for investment advisory services and carries no proprietary trading book: no commissions on investment transactions, no proprietary positions, and no incentive to trade against the client.
The honest limit matters just as much as the indictment. The public enforcement record documents spoofing, attempted manipulation, benchmark abuse, false statements, and supervisory failure. It does not, in publicly verified sources, document a clean, proven pattern of banks simultaneously driving prices down while placing large physical delivery orders. Some market participants argue that broader thesis. The narrower claim is the documented one: the same banks at the center of paper price discovery have been fined repeatedly for manipulating those markets.[14][16][17][18][19][20][21][22] When a theory outruns the receipts, say so plainly.
And that returns you to the larger thesis of this essay. A paper market that can be spoofed, fined, redesigned, and modernized by committee is exactly why the physical ounce matters. The ounce cannot be redefined by benchmark governance and it cannot be canceled by a trader’s keystroke. In that sense, the physical bar remains the honest ruler. PHYS and PSLV exist precisely because the paper tape has a rap sheet.
Reserves Are a Disclosure, Not an Inventory
Official gold reserves present a similar problem.
IMF reserve statistics generally depend on information reported by national authorities. They record monetary gold held as an official reserve asset, not every ounce held by a sovereign wealth fund, state institution, or government-linked entity.[28]
The World Gold Council’s country reserve data draw heavily on IMF statistics but may incorporate other information. Its central-bank demand estimates are a flow measure and can differ from changes in reported reserve stocks because of reporting lags, classification choices, other official buyers, and purchases that are disclosed later.[29]
Those discrepancies are documented. The explanation is not always provable.
It is reasonable to say that a country may delay reporting purchases or hold gold outside the narrow reserve category. It is not reasonable to treat every estimate of hidden Chinese gold as established fact. China’s reporting history has produced gaps and sudden additions, but the exact amount of any unreported stock is unknowable from public data.
This is where prudence matters. A reserve manager should treat reported gold as a disclosure, not as a complete physical inventory. A household should treat a brokerage account statement as a record of legal claims, not as a warehouse inspection.
What Gold Is Actually Saying
The structural case for gold does not require a permanent bank-set ceiling, a single dramatic “COMEX break,” or a cinematic theory of suppression.
The documented facts are more interesting.
World Gold Council data show unusually strong official-sector gold buying after 2021: approximately 1,082 tonnes in 2022, 1,037 tonnes in 2023, about 1,044 tonnes in 2024, and 863 tonnes in 2025, subject to revisions and differences between reporting vintages.[30]
The timing matters. The 2022 acceleration followed the freezing of Russia’s foreign-exchange reserves. Gold became a reserve asset less directly exposed to the seizure of foreign-currency claims. That does not mean gold became invulnerable or literally “sanction-proof.” It means reserve managers were reminded that a foreign asset can be legally immobilized even when it is listed on a balance sheet.
BRICS is the political face of this shift, not necessarily the daily mechanical driver of every gold tick. The structural bid is official central-bank accumulation. Political blocs provide the narrative and strategic context; they do not dictate each intraday auction.
The more useful formulation is not “the dollar will collapse.”
It is:
Gold is rising in every fiat currency.
The dollar may remain the world’s principal invoice currency while gold rises in dollars, euros, yen, and yuan. DXY can sit near 100 while gold reaches new highs. Those facts are not contradictions. They mean gold is being repriced as a reserve asset against multiple paper currencies at once.[30]
Silver adds another layer. Reports of depleted visible Shanghai inventories, local premiums, and episodes of backwardation point to physical tightness in China. Some market commentary has described visible silver stocks as roughly 90 percent below their 2020–21 peaks, but those figures combine exchange-specific measures and should not be treated as one perfectly consolidated inventory series.[31]
The proper conclusion is not that BRICS has “broken COMEX.” It is that Chinese physical tightness can lead Western paper markets. COMEX and London futures can cap a week, accelerate a liquidation, or delay recognition of scarcity. But when physical premiums, withdrawals, lease rates, backwardation, and delivery demand begin pointing in the same direction, paper markets may lag and then reprice violently.
Open-interest declines are also documented in periods of precious-metals volatility. But falling open interest does not prove manipulation. It can reflect liquidation, contract rolls, reduced leverage, or a market in which speculative participation has been washed out.[32]
The disciplined statement is therefore narrower:
- COMEX and LBMA remain important venues for price discovery.
- Physical tightness can create a lag between regional metal demand and Western paper pricing.
- Open-interest collapse can make a market thinner and more volatile.
- None of those facts proves a permanent bank-set price ceiling.
- None proves that one dramatic event “broke” a futures exchange.
Reality rarely offers the satisfaction of a single villain.
What This Means for a Metals Framework
For a diversified metals sleeve, the first distinction is between ounces and mining equities.
Physical-ounce vehicles such as PHYS and PSLV represent exposure to metal-related structures; miners such as GDX, GDXJ, SIL, and SILJ are equities. Miners carry operating costs, political risk, financing risk, dilution risk, management risk, and the risk that a rising gold price is swallowed by rising costs.
The internal Regatta sleeve framework places the physical-metals sleeve within a 10–16 percent band, with a default 60/40 gold-to-silver split while the gold-silver ratio remains in the 60–70 range. It contemplates moving toward 70/30 if the ratio holds above 72 for two weekly closes, toward 50/50 only if the ratio breaks below 58 with physical-tightness confirmation, and capping PSLV at 55 percent of the metals sleeve during January-style spikes.[33]
These are educational framework parameters, not individualized advice.
The more important principle is the miner kill-switch:
Is an ounce buying more diesel, labor, and steel than it did 12 months ago?
If the answer is no, official CPI may be lagging the costs that matter to miners. A mining equity can look cheap relative to gold while its operating leverage is quietly deteriorating. The intermediate processed and unprocessed PPI series are more relevant to that question than a calm final-demand headline.
In that environment, the framework favors the ounce sleeve over pretending that every miner is a leveraged ounce.
That is Essentialism applied to markets: reduce the decision to the few variables that actually govern the outcome. Not every chart deserves a vote.
The Practical Rule
When a report says gold is up or down since 1980, ask which inflation series was used.
When it says the dollar is strong, ask: DXY or gold?
When it says money supply exploded, ask whether the definition changed.
When it says inventories crashed, ask: registered, eligible, or total?
When it says inflation is contained, ask which PPI: the final-demand blend or the production pipeline?
When it says central banks bought gold, ask whether the figure measures reported reserve stocks, estimated official-sector flows, or both.
The point is not to reject official statistics. It is to read them with responsibility.
A ruler is useful only when you know what it measures.
The modern statistical system is not a rose-colored conspiracy. It is a collection of changing instruments, each built for a particular purpose. The trouble begins when a household uses an imputed shelter measure to understand its mortgage, a reserve manager uses DXY to understand global purchasing power, or an investor uses final demand to understand mining costs.
The formula moved.
The metal did not.
Sources
[1] U.S. Bureau of Labor Statistics, Producer Price Index release for August 2026
[2] BLS PPI data and detailed tables
[3] BLS, “Owners’ Equivalent Rent and the Consumer Price Index”
[4] National Academies, discussion of CPI measurement and the Boskin Commission
[5] BLS, CPI historical changes and methodology
[6] BLS, Producer Price Index history
[7] BLS, “The PPI’s Final Demand–Intermediate Demand System”
[8] Federal Reserve, discontinuation of M3
[9] Federal Reserve, “An Update to Measuring the U.S. Monetary Aggregates”
[10] ICE, FX Indexes Methodology
[11] LBMA, precious-metals benchmark history and methodology
[12] LBMA, Good Delivery technical specifications
[13] CME Group, metals delivery and warehouse-stock information
[25] Reuters, “Moore Capital settles platinum, palladium lawsuit for $48.4 mln”
[26] Gold litigation settlement filings and allocation materials
[28] IMF, gold in central-bank reserves
[29] World Gold Council, gold reserves by country and methodology
[30] World Gold Council, central-bank gold demand data
[31] Shanghai silver inventory reporting and market data
[32] CFTC, COMEX futures positioning and open-interest data
[33] Regatta Financial internal metals-sleeve note, September 9, 2026. Internal educational framework; not a public source or individualized investment recommendation.
Disclosure
This article is provided by Regatta Financial LLC for educational and informational purposes only. It is not investment, tax, legal, accounting, or individualized financial advice, and it is not a recommendation to buy or sell any security, fund, commodity, or other financial instrument. References to PHYS, PSLV, GDX, GDXJ, SIL, SILJ, COMEX, LBMA, ICE, the Federal Reserve, the IMF, BLS, and other named entities are for analytical and educational purposes only. Precious metals, mining equities, exchange-traded products, and futures-related investments involve risk, including volatility, liquidity, operational, counterparty, currency, political, and loss-of-principal risks. Past performance is not indicative of future results. Readers should consult qualified professionals regarding their individual circumstances.
Leave a Reply