Over roughly two centuries, the silver in Rome's everyday coin fell from about 95 percent to about 5 percent. No single emperor destroyed it. Each one made a small, defensible decision in a difficult year. The people holding those coins were the last to find out.
Most people know how Rome ends. A date, 476, and a barbarian king deposing a teenage emperor. It makes a tidy story and it teaches almost nothing, because by 476 the outcome had been settled for generations. The empire did not fall in 476. It had been thinning for two hundred years, and the thinning is the part worth studying.
Consider a Roman shopkeeper in the year 150. Trade functions. The roads are maintained. Aqueducts run. Construction continues in his city. He holds a denarius and it looks exactly like the denarius his grandfather held: same size, same weight in the palm, same emperor's profile, same silver shine. He has no reason to believe anything is wrong, and by every measure available to him, nothing is.
But that coin is about 75 percent silver, and his grandfather's was 95. Somewhere in the intervening decades, a series of reasonable men facing difficult years made a series of small adjustments. None of them announced it. Most of them were solving a real problem: an army to pay, a fire to rebuild after, a frontier that would not hold. Each adjustment was defensible on its own. The sum of them was not.
The Romans developed surface-enrichment techniques. They treated the debased coin so the outside carried more silver than the inside. The coin looked right. It rang right. On the newer issues, the shine wore off with handling and the copper showed through underneath.
They were not merely debasing the money. They were engineering the appearance of stability while the substance left. That is a different and more troubling act, and it is the one that has a modern echo.
I want to be careful about what I am claiming here, because this comparison is made carelessly all the time. I am not going to tell you that America is Rome, that a date is coming, or that history repeats on a schedule. It does not, and anyone selling you that certainty is selling you something. What I am going to do is lay the two records side by side and let you judge which of the mechanisms look familiar.
Each pair below sets one Roman mechanism against its modern counterpart. Read them as mechanisms, not prophecies. The question is not whether we are Rome. The question is whether the machinery is the same machinery, and whether it is pointed the same direction.
Nero began it after the fire of AD 64 with the cost of rebuilding a capital in front of him. Trajan continued it. Marcus Aurelius, the philosopher-emperor whose private notebooks are still read as a manual for personal integrity, presided over a slide to roughly 75 percent. Septimius Severus took it to about half. Caracalla issued a new coin marked at two denarii that carried the silver of about one and a half.
Not one of these men set out to destroy Roman money. Each faced an army that had to be paid this year with revenue that did not exist this year.
Measured by the Bureau of Labor Statistics' own consumer price series, the dollar has lost roughly 97 percent of its purchasing power since the Federal Reserve opened in 1913. What one dollar bought then takes about thirty-three today.
The curve is not uniform. It is gentle until 1971, when the last convertibility to gold ended, and steeper afterward. A dollar held in cash since January 2020 has lost roughly a fifth of its purchasing power in six years. As in Rome, no single decision did this, and every individual decision had a defender.
Roman expansion under the Republic had been self-funding: conquest paid for the army that made conquest possible. Once the borders stopped moving outward, that engine reversed. The empire now had to pay for what it already held, out of what it already produced.
Army pay rose repeatedly. Emperors bought loyalty with donatives on accession. The grain dole fed the capital. When revenue would not cover it, emperors reached for the coinage rather than the tax rolls, because debasement is quiet and taxation is loud.
Federal debt stands near 39.8 trillion dollars as of late July 2026. Debt held by the public is around 101 percent of GDP this year and is projected by the Congressional Budget Office to reach 120 percent by 2036, passing the 1946 wartime peak of 106 percent.
The deficit for fiscal 2026 is roughly 1.9 trillion dollars, about 5.8 percent of GDP, against a fifty-year average near 3.8 percent. This is being run at 4.2 percent unemployment with the economy growing. It is not a recession deficit.
A rising share of Roman output went to holding position rather than improving it: standing armies on static frontiers, an enlarged administration, and in some periods outright payments to tribes in exchange for not invading. Diocletian's reforms brought stability at the price of a substantially larger army and bureaucracy, which had to be fed by a heavier and more coercive tax system.
Late imperial legislation increasingly bound farmers to their land and sons to their fathers' trades. That is what a state looks like when it can no longer afford the flexibility it used to permit.
Interest on the federal debt reached about 970 billion dollars in 2025 and is projected near 1.04 trillion for fiscal 2026, roughly 3.3 percent of GDP. That is above the post-war record of 3.2 percent set in 1991. It consumes an estimated 18.6 percent of all federal revenue this year.
Interest already exceeds Medicare. CBO projections have it passing defense and non-defense discretionary spending by 2038 and becoming the single largest line in the federal budget by 2048. Interest buys nothing. It is rent on decisions already made.
In AD 301 Diocletian issued the Edict on Maximum Prices, capping more than 1,200 goods, services and wages across the empire, with death prescribed for violations. Its preamble did not identify the coinage as the cause. It identified avaritia, the greed of merchants and speculators.
The result was recorded by Lactantius: sellers withdrew goods rather than trade at a loss, and scarcity grew worse than before. Trade moved to the black market. Diocletian eventually abdicated and, by his own account, took up growing cabbages.
The reflex to locate inflation in seller behaviour rather than in monetary and fiscal conditions is not confined to one political tradition. Both American parties have reached for price caps, anti-gouging statutes and rent restrictions within the past decade, and the rhetorical framing has been remarkably consistent across them.
Economists genuinely disagree about how much of the 2021 to 2023 episode ran through corporate margins. What is far less disputed is that a cap does not create supply, and that shortages appear wherever the cap binds hardest.
The city of Rome ate Egyptian grain. The annona, the grain supply that fed the capital and kept it politically quiet, depended heavily on a single province and the sailing season that connected it. Emperors understood this well enough that Egypt was governed as the emperor's personal preserve and senators were barred from entering it without permission.
Rome also came to rely on federated Germanic forces to hold frontiers against other Germanic forces. Both arrangements worked for a long time. Both meant the survival of the centre rested on something the centre did not control.
This is the pattern I spend most of my working life on, though usually at household scale rather than imperial scale. A utility family will often have salary, pension, health coverage in retirement and the single largest holding in the portfolio all resting on the same employer.
Four loads, one source, no second path. It works for a long time, and the years it works are exactly what makes it feel safe.
Between AD 235 and 284 the empire ran through somewhere around two dozen recognised emperors, plus a longer list of claimants who did not last long enough to be counted. The typical reign was measured in a couple of years and ended violently.
In that environment, debasing the coinage to pay the legions this quarter was not stupidity. It was the rational move for a man whose planning horizon was however long the army stayed loyal. The behaviour that destroyed the currency over two centuries was, for each individual emperor, obviously correct.
Ours are gentler versions of the same structure. Congressional terms run two and six years. A president has four. A Fed chair has four. Corporate management reports quarterly. Very few of the people making decisions with thirty-year consequences will be in the seat in thirty years.
The debt problem is not principally a problem of bad people. It is a problem of a system in which nobody's term is as long as the liability.
A page like this is only worth reading if it tells you where it is weak. The Rome analogy is one of the most abused arguments in financial commentary, and I would rather hand you the objections than have you find them later and discount everything else I have written.
The historian Alexander Demandt catalogued 210 separate published theories for Rome's fall, and more have been added since. Anyone who tells you it was the money is choosing one thread from a very large tapestry.
Numismatists including Kevin Butcher have argued that debasement was largely a symptom rather than a driver: Rome debased because plague, invasion, civil war and a tax system that could not reach its own provinces had already broken the state's finances. On that reading, the coinage was the thermometer, not the fever.
Removing silver from a coin that circulates because it contains silver is a breach of a specific representation. A fiat currency makes a weaker and more honest promise, and every holder knows it. These are not the same act, and treating them as identical is the analogy's biggest single weakness.
Constantine's gold solidus held its standard for centuries and outlived the Western empire by roughly a thousand years in the East. The Roman monetary story does not only demonstrate that money decays. It demonstrates that decay was reversible when a state chose to reverse it.
The United States borrows in the currency it issues, holds the world's reserve currency, and still clears its auctions with real demand. As of early 2026 bid-to-cover ratios ran roughly 2.3 to 2.9 across bills, notes and bonds. Rome had no equivalent advantages, and no productivity growth engine to grow out of anything.
Rome's slide took roughly two hundred years. Even a reader who accepts every parallel above should notice that the timescale offers no guidance whatsoever about the next five years, and that treating it as if it does is the error that turns history into a sales pitch.
Not that collapse is coming. Not that a date can be named. I have written on this site before that I do not have a view about the quarter, and that has not changed.
What I claim is narrower and, I think, harder to argue with. The reference period that most retirement planning rests on runs roughly from 1945 to now. It is a single sample from an unusually favourable stretch of history: American industrial dominance, a young workforce, cheap energy, forty years of falling interest rates from 1981, and a debt burden that started near wartime peak and fell for three decades before turning. Nearly every rule of thumb a retiree is handed was fitted to that period.
Rome's value to us is not as a prophecy. It is as evidence that long stretches of apparent normality can sit on top of a slow deterioration, and that the people living inside them are not equipped to see it from the inside. I do believe we are living through a significant repricing. I hold that view openly and I could be wrong about it.
Here is where I have to be precise, because this is the point at which historical commentary usually turns into something worse. The argument I am making is not that the standard principles of investing are wrong. Most of them are not. The argument is that a number of the specific rules retirees are handed are not principles at all. They are calibrations, fitted to a single unusually favourable eighty-year sample of one country's history, and presented as if they were laws.
The distinction matters because it points in a direction people find counterintuitive. If the reference period is less reliable than advertised, the correct response is more diversification, not less. It is not a reason to concentrate into whatever you believe will win, and it is emphatically not a reason to try to time an exit. A Roman who correctly understood in AD 150 that the coinage was decaying would still have been wrong about almost every date he might have picked.
Do not let one source carry your income, your health coverage and your largest holding at the same time.
Costs, fees and taxes compound against you in every regime, and they are the one input you control outright.
Match the asset to the liability and the horizon. Money needed in three years and money needed in twenty-five are not the same money.
How you behave in a drawdown matters more than the allocation you chose before it.
Liquidity has an option value that looks like waste until the moment it does not.
That a 60/40 split is the neutral default. It depends on stocks and bonds moving in opposite directions, which held through much of 2000 to 2020 and conspicuously failed in 2022 and through the 1970s.
That bonds are the safe part. They are safe against equity risk. They are not safe against inflation, as anyone holding long Treasuries from 2020 to 2023 discovered in real terms.
That stocks always recover within a tolerable window. True of the US sample. Japan's 1989 high took more than thirty years to regain in nominal terms.
That the withdrawal-rate conventions are laws. They were derived from one country, one currency, one historical run.
That buying the index is inherently diversifying. The index today carries more concentration in a handful of names than at almost any prior point in its history.
None of the above is investment advice, and none of it is a recommendation to buy or sell anything. What is right for a particular household depends on facts I do not know from a web page: your pension election, your tax position, your health, your spending, your timeline and your actual tolerance for a bad year. Two people reading this paragraph should reasonably end up in different portfolios.
If the page has any practical value, it is here. These are the five questions I would want answered about my own situation before deciding whether any of the foregoing changes anything.
If my employer had a genuinely bad decade, how many of my sources of security would move at the same time? Count them honestly: salary, pension, retiree health coverage, and the largest position in the portfolio.
What historical period was my plan's assumed rate of return measured over, and what specifically would have to remain true for that assumption to keep holding?
Which part of my portfolio protects me against a fall in stock prices, and which part protects me against a loss of purchasing power? If the answer to both is the same holding, that is worth examining.
If I had to fund the next three years of spending without selling anything at a loss, could I do it? This is the question that determines whether a bad market is an inconvenience or a permanent impairment.
Who in my financial life holds a horizon as long as mine, and how exactly are they compensated? The Roman emperors debased the coinage because their planning horizon was shorter than the consequence. That structure did not disappear.
The Roman in AD 150 had no way to assay the metal in his hand. He could not read a monthly Treasury statement or a CBO projection. He had no access to any of the instruments that would have told him what was happening to his savings, and so he did the only rational thing available and carried on.
We do not have that excuse. The debt figures are published. The interest line is published. The concentration in your own portfolio is a number somebody can put in front of you this week. The disadvantage the shopkeeper had was information. Ours is attention.
If you want a straight assessment of how much of your retirement rests on a single source, and what the plan actually assumes about the next twenty years, that is the conversation I have most weeks.