Real vs. nominal

The most consequential idea in finance is also the most easily missed: nominal and real are not the same thing. The nominal number is the literal dollar amount on the screen — the price of a share, the figure on a paycheck, the line in a brokerage statement. The real number is what those dollars can actually buy. They are almost never identical, and over long stretches of time the gap between them swallows entire fortunes. Most people in their lifetime experience growth that is partially or fully an illusion of the unit of measure — and they confuse that illusion with the substance.

The most-quoted statistic in American investing is that the S&P 500 went from about 100 in 1971 to roughly 5,900 in 2024. That is a 59× return — a five-thousand-nine-hundred-percent gain — and it is true. It is also, in any meaningful sense, misleading. Adjust the same series for the Consumer Price Index and the 59× shrinks to 8.9×. Adjust it for the broad money supply (M2) and it shrinks to roughly 1.9×. Price the index in ounces of gold and the entire fifty-three-year run effectively round-trips — an ounce of gold buys about as many index points today as it did the day Nixon closed the gold window. Three deflators, three answers, all from the same data. The chart you choose to look at is a political act before it is a mathematical one.

The same trick applies to wages, and the result there is more painful. US median household income rose from $9,870 in 1971 to about $80,610 in 2023 — an 8.2× nominal raise that looks, on its face, like a workers' bonanza. Convert it to real CPI dollars and the gain is 27% over fifty-two years, a rate of roughly half a percent per year. But price the median paycheck in the assets the median household actually needs to retire on — a home, a college degree, a healthcare plan, a stake in the stock market — and the picture inverts. Income measured in median home prices has lost roughly half its purchasing power since 1971. Income in healthcare units has lost more than half. Income in college tuition has lost roughly sixty percent. Income in shares of the S&P 500 has lost ninety percent. The median worker has, by the only measures that actually matter for life outcomes, gotten poorer over five decades — even while their paycheck has gone up eight-fold.

The honest version of this requires admitting two uncomfortable things about the official inflation statistics. First, the CPI's basket is fixed, with hedonic adjustments, owners' equivalent rent in place of actual house prices, and a steady drift toward measuring the things that are getting cheaper rather than the things that are getting more expensive. Even the official real numbers — the ones that flatter the policy narrative — are too flattering. Second, the only deflator that captures the full scale of asset-price inflation is broad money supply: M2 in the United States, plus Eurodollars and shadow-bank money creation everywhere else. When you measure asset returns against the actual quantity of money that exists, most asset returns turn out to be roughly ordinary. The S&P beats M2 by perhaps a percent or two a year over very long stretches. The 59× becomes 2×. The miracle becomes mathematics.

This is also why the famous Jeremy Siegel argument — "stocks for the long run, twenty years is always enough" — is technically true and practically misleading. Yes, an investor who bought a diversified equity index in 1900 and held to 2024 would have compounded magnificently in dollars. But the same investor over 1968–1988 — a textbook "twenty-year horizon" — earned a 4× nominal return, only 0.97× in CPI-adjusted real terms, and about 0.15× in gold terms. Two decades of disciplined investing produced a real loss against the official deflator and a catastrophic loss against hard assets. The investor who held from 2000 to 2013 lost in nominal dollars, never mind real ones. Long-run averages are statements about the survivors of history, not promises to the people living through it.

What this means for actual portfolios is simple to state and rarely practiced. The disciplined investor thinks in real terms even when reporting nominal ones. The relevant benchmark for a retirement portfolio is not "the S&P 500 in dollars." It is the real liability the portfolio exists to meet — a stream of retirement income measured in goods you will need to buy, the cost of a home for your kids, the cost of an education for them, a stream of healthcare in your old age. A portfolio that "beats the S&P" in dollars but loses ground to median home prices is not actually doing its job. It is performing for a metric that does not feed you.

The same logic explains why the gold price, the Bitcoin price, and the price of urban real estate persistently set new all-time highs in dollars — not because those assets have become individually more valuable, but because the dollar has become individually less valuable. The denominator is shrinking faster than the numerator is growing. Headlines about "all-time highs" should be read with the same skepticism one would apply to a yardstick that quietly got shorter every year and was then used to measure how much taller the children had become. The children are taller. They are not as much taller as the yardstick says.

There is one final and harder point. Even when an asset does compound in real terms, the compounding accrues to the owners of capital, not to the earners of wages. The S&P's 8.9× real return is real for the household that owned a meaningful stake in it for fifty years. For the household that did not — which is roughly the bottom half of American households by wealth — the real benchmark is the assets they failed to acquire, and against that benchmark the same fifty years have been a period of relative impoverishment. Real-vs-nominal is not just a technical correction. It is a redistribution map. The dollars that quietly lose purchasing power leave one pocket; the assets that quietly absorb that purchasing power sit in another. The transfer is silent, continuous, and visible only when you change the unit of measure.

The instrument above gives you the lens to do that change. Toggle the deflator on the S&P. Re-price assets in gold, in money supply, in hours of minimum-wage labor, in big macs. Look at the chart of your favorite asset against the unit you actually care about — your salary, your rent, your tuition bill. Once you have seen the same data through five different units, the headline number on any single chart will never feel quite the same. That is the lesson. Not that any one deflator is correct, but that no single deflator is sufficient.