On August 12 the Bureau of Labor Statistics reported that consumer prices rose 0.1 percent in July and 3.4 percent over the previous twelve months, a tenth of a point below June. Markets treated the print as confirmation that the measuring rod is behaving.
Now measure the same economy with a different rod. On August 17 the Dow Jones Industrial Average closed at 53,459.78, near its all-time high. Gold was trading around $4,400 an ounce. Divide the first number by the second and the Dow costs about twelve ounces of gold. In early 2024 it cost about nineteen. Priced in gold rather than in paper, the American stock market has lost roughly a third of its value in two and a half years—over exactly the stretch in which it kept setting nominal records.
Both descriptions are accurate. They are the same market expressed in two different units. Everything interesting in monetary economics lives in the gap between them.
Every Price Is a Ratio, and Both Sides Move
A monetary price is an exchange ratio between goods and money. When the ratio changes, something happens on one side or the other, and the number itself does not tell you which. The consumer price index resolves that ambiguity by assumption: it treats the money side as the fixed reference and books every movement against the goods.
That assumption is not a technical detail. It is the entire content of the statistic. If the dollar is the yardstick, then the yardstick cannot be short. A measuring system built on the currency is structurally incapable of registering what happens to the currency.
Austrians have made this point since Mises, and it is usually dismissed as pedantry. It stops being pedantry the moment you notice that the yardstick has an owner, and the owner has policy objectives.
The Rod Gets Rebuilt, and Always in the Same Direction
Grant the framework anyway, and a second problem appears: the index is not even a fixed rod in its own terms. It is a maintained statistical product, redefined periodically by the agency that publishes it.
In January 1983 the BLS stopped pricing owner-occupied housing by what houses cost and switched to rental equivalence—an estimate of what an owner would hypothetically pay to rent his own house. In 1996 the Boskin Commission concluded that the index overstated inflation by about 1.1 percentage points a year. The BLS then adopted geometric-mean formulas at the lower level of aggregation and steadily broadened hedonic quality adjustment, which discounts a price increase to the extent the product is judged better than the one it replaced.
Each change had a serious technical rationale, and I am not alleging fraud. I am pointing at a pattern. Every major revision of the past forty years has lowered measured inflation relative to the method it replaced, and every one was adopted by the institution whose fiscal obligations—Social Security, tax brackets, indexed debt—are escalated by the resulting number.
You can see the machinery in the July report itself. Shelter accounted for roughly two-thirds of the monthly increase, and the largest single component of shelter is not a price anyone paid. It is imputed rent on houses that are not for rent.
A Rod No Committee Maintains
Gold has none of this. There is no methodology board, no seasonal adjustment, no annual reweighting, no revision window. Not because gold has a constant value—it plainly does not—but because nobody owns the definition. An ounce in 1932 and an ounce today are the same object.
That is the only property required for the job. To audit a currency you need a reference the currency’s issuer does not control. Denominate the Dow in ounces and you get a series that no institution administers, adjusts, or has an interest in.
What the Unmanaged Rod Shows
The historical record of that series is remarkably clean. It swings between extremes, and the extremes mark the turning points of the twentieth century.
In September 1929 the Dow was worth about eighteen ounces. By July 1932 it was worth two. In February 1966 the Dow reached 995 while gold sat at $35 by statute—about twenty-eight ounces. By January 1980, with the Dow near 875 and gold spiking to $850, it was worth roughly one. In August 1999 the Dow crossed 11,300 against gold near $255: more than forty ounces, the highest reading on record. By 2011 it was back to about six.
Then the sequence broke. From 2011 to 2024 the ratio drifted sideways and upward for twelve years, an unprecedented pause, while the Federal Reserve ran successive rounds of asset purchases and corporations bought back their own shares with the proceeds. The pause ended in February 2024. Since then the ratio has fallen from about nineteen to about twelve.
Note what this means for the ordinary investor. In dollars, his account is at a record. In purchasing power over the one asset no central bank can print, he has given back a third of it since 2024. No brokerage statement reports the second number, and no inflation release will ever contain it.
Why the Ratio, and Not the Price Index, Is the Cycle Instrument
This is where Austrian capital theory does real work rather than decorative work. Newly-created credit does not raise all prices at once and in proportion. It enters at specific points, and it enters first where the credit itself goes: into long-duration assets whose valuations hang on a discount rate. Equities, real estate, long bonds, and the ventures that only pencil out at low rates all inflate long before the effect reaches the supermarket.
A consumer price index is therefore not merely an imperfect instrument for detecting monetary expansion. It is pointed at the wrong place. It measures the last stage of a process whose first stage is the whole story.
Gold moves opposite that process because gold is the one financial asset that is nobody’s liability. It sits out the boom, which is why it looks like dead money for twenty years at a stretch, and it does the repricing when the boom unwinds. The Dow/Gold ratio captures both halves in a single number: the numerator is the credit cycle, the denominator is the exit from it. The same four decades that produced those swings also produced a long decline in labor’s share of non-farm business output—the same redistribution, recorded by a different instrument.
A Number That Can Be Wrong
I will state the claim in a form that can fail, since almost nothing in macroeconomic commentary is stated that way.
The lows of the completed cycles fall in a straight line: about two ounces in 1932, about one in 1980. Roughly half a century apart, and each one half the last. Extend it and the current cycle terminates near half an ounce, some time around 2030. That is the target I defend in my book, and it is deliberately specific.
It is also falsifiable. If the ratio turns up from twelve and exceeds the 1999 high above forty without first reaching single digits, the thesis is finished—not weakened, finished. There is no reweighting available to me, no substitution effect to invoke, no revision window. That is the price of using a rod somebody else cannot rebuild, and it is a price worth paying.
The July CPI report told us the rod is behaving. It could not have told us anything else. To learn what the rod is made of, you have to measure it against something no committee maintains.
