The World Is Supposed to Be Getting Cheaper

A long-distance phone call to London used to cost a dollar a minute. Today it's free, in better fidelity.

A set of encyclopedias used to cost thousands of dollars and take up a shelf. Today the same reference work lives in your pocket, free, and updates itself overnight.

A roll of film used to cost what a meal cost, plus a trip to the drugstore to develop it. Now you take a thousand photos for nothing and don't think twice.

Pick almost any product where the cost is dominated by information or coordination, and the same pattern shows up: marginal cost falling toward zero, year after year, decade after decade. Productivity gains and digital reproduction have made the world, in real terms, extraordinarily cheaper than it was forty years ago.

And yet the cost of living keeps going up.

That's the puzzle this article unwinds. A separate piece covered the dollar-erosion side of the ledger — same bill, less stuff. This one covers the other side: the suppressed deflation you should have been feeling but weren't. In a sound-money world — one where the supply of dollars isn't expanded year after year — your paycheck would compound with productivity. Each year the same number would buy more. That's not the world we live in. The dollar got cheaper faster than the technology did, and the productivity gains got absorbed into prices instead of paid back to you.

Productivity deflation, named

Before going further, name the thing.

When it takes fewer labor-hours to produce a good, the price of that good should fall — if the measuring stick is held still. If a factory that built one car in a hundred hours now builds the same car in fifty, the cost of producing the car has been cut roughly in half. In a stable monetary environment, that cost cut shows up as a falling sticker price. The household buying the car gets the productivity gain.

That's productivity deflation, and it's not an obscure theory. It's roughly what the late-19th-century United States looked like. From roughly 1873 to 1896 — a period later called the “Great Deflation” — US wholesale prices fell on average about 1.5% to 2% per year while real output grew and real wages rose. The economy industrialized and got dramatically richer in real terms. Prices, on average, went down.

Most readers' intuition about deflation comes from a different episode: the 1930s. Collapsing demand and a banking crisis produced a falling price level alongside mass unemployment.

But the 1930s deflation ran on a different mechanism — debt unwinding and demand collapse, not productivity surplus. Treating the two as the same thing is a category error, and telling the dangerous kind of deflation from the harmless kind is a subject of its own.

Andrew Atkeson and Patrick Kehoe surveyed 17 countries and 100 years of data in “Deflation and Depression: Is There an Empirical Link?” (2004). Out of 73 deflationary episodes they identified, 65 weren't associated with depression at all. Almost every case outside the 1930s breaks the textbook intuition.

This article is about the productivity kind. The kind that should have shown up on your receipts and didn't.

The receipts: what genuinely got cheaper

Information goods are the clean case. The marginal cost of one more copy of a digital file is essentially zero, so as soon as the distribution medium became digital, the price had nowhere to go but down.

Long-distance calls. In 1980, AT&T charged in the neighborhood of a dollar a minute for a peak-time international call to London. A ten-minute call cost more than a paperback book. Domestic long-distance was cheaper but still metered, in the range of twenty to forty cents per minute. Today the same ten-minute call to London, in better fidelity, is free over WhatsApp, FaceTime, or any similar app.

The encyclopedia. The 1990 Encyclopædia Britannica 32-volume set listed at about $1,400 — a multi-thousand-dollar reference work that took up a shelf and went out of date the day it shipped. By 2010 the print edition was discontinued. Wikipedia now has roughly fifty times the article count, costs nothing to read, and updates continuously. The function (look something up, authoritatively) didn't disappear. The price of the function did.

Photography. A 24-exposure roll of film plus drugstore developing ran ten to fifteen dollars in the late 1990s — about forty cents per shot, before you'd seen any of them. The same phone in your pocket now takes effectively unlimited photographs at zero marginal cost, with better dynamic range and instant sharing.

Turn-by-turn navigation. Garmin's StreetPilot, the first portable mapping GPS (Global Positioning System) for cars, launched in 1998 at about $400 — celebrated at the time as the first practical, affordable consumer navigator. Within a decade, comparable standalone units were under $100. By 2015, Google Maps and Apple Maps were free, bundled with the phone you already owned, and better than anything a dedicated unit shipped.

Music, video calls, movie rentals, classified ads, software distribution, photo storage — the same shape, case after case. Pick a category where the cost was dominated by physical reproduction or middleman coordination, and the marginal cost has been falling toward zero for thirty years.

That's the deflation side of the ledger. Now the other side.

And the price level went up anyway

If the productive part of the economy was getting radically cheaper, you'd expect the aggregate price level to be flat at minimum — and probably falling. That's not what the receipts show.

The Consumer Price Index, which tracks a representative basket of household goods and services, sat near 78 in January 1980. It's above 330 today. The price level, on average, has more than quadrupled. That figure includes the radically cheapened information goods and everything else, weighted by spending. The weighting matters.

Look at what got more expensive over the same window:

• Rent has more than quintupled in nominal terms. The median gross rent in the US (Census figure, includes utilities) ran about $243 a month in 1980 and clears $1,400 today.
• College tuition at private four-year institutions is up roughly twelve-fold. A degree that cost about $3,500 per year in 1980 now lists at $45,000 or more.
• Medical care is up roughly seven-and-a-half-fold, comfortably ahead of overall CPI.

You can substitute toward Wikipedia. You can't substitute toward a Wikipedia of college, or a free streaming version of your kid's pediatrician.

So the receipts tell a forked story. The information-goods side has been in deflation for thirty years. The household-essentials side has been in heavy inflation for the same thirty years. The aggregate price level went up because the essentials are weighted heavily in the basket and rose faster than the cheap stuff fell.

The puzzle isn't whether productivity made things cheaper. It demonstrably did. The puzzle is why so little of that gain reached the household budget.

What ate the productivity gains

The dollar shrank faster than the technology cheapened.

Most new dollars are created when commercial banks lend, a mechanism walked through in detail in another article. The broad money supply — the Federal Reserve's M2 measure — grew from about $1.5 trillion in 1980 to roughly $22 trillion today. That's more than a fourteen-fold expansion against real output that roughly tripled over the same window. Put plainly: the supply of dollars grew much faster than the supply of real stuff to buy with them. Over decades, that mismatch shows up as a rising average price level.

In a stable monetary environment, productivity gains would land as falling sticker prices — the car factory from earlier, applied to the whole economy. That's not what happened. In productive sectors, sticker prices mostly stayed flat (your laptop costs roughly what a laptop cost fifteen years ago, even though it's vastly faster). Everywhere else, sticker prices rose. The gains were absorbed.

This is not an accident. The Federal Reserve adopted an explicit 2% inflation target in 2012, formalizing a policy stance that had been operating informally for decades. The Fed's target is roughly 2% inflation a year. The Fed itself defends the target as a buffer against deflation risk and a tool for managing employment. Whatever the rationale, the effect is the same: productivity gains aren't allowed to show up fully as falling prices — when technology makes things cheaper to produce, monetary expansion offsets part of that decline. Strip the terminology and what the target effectively does is this: when your laptop should have gotten cheaper, the system makes sure on average it doesn't.

Jeff Booth, in The Price of Tomorrow (2020), made this the spine of his argument: the natural state of a free market with compounding productivity is falling prices, and the policy commitment to ~2% inflation is specifically designed to defeat that. You don't have to agree with every claim in the book to see that the structural setup is what he describes.

The Cantillon angle

The productivity gains weren't lost. They were redistributed.

When new money enters the economy through bank lending and Fed balance-sheet expansion, it doesn't arrive evenly. It arrives at specific points — large banks, big borrowers, asset markets — and ripples outward. The people closest to the money source benefit before prices catch up. The people farthest away (wage earners, savers) benefit later, or not at all.

The mechanism shows up cleanly in the divergence between asset prices and wages. From March 2009 to early 2022 — the years of post-2008 Fed balance-sheet expansion — the S&P 500 rose roughly 600%. Real wages, adjusted for inflation, rose less than 15% over the same window. Asset prices grew about forty times faster than wages.

A household that owned assets — stocks, real estate — rode the asset-price expansion the new money produced. A household that earned its money from wages mostly didn't. That's the Cantillon effect, and it's covered in detail elsewhere. The point here is just that the missing productivity dividend is not actually missing. It went somewhere specific.

The world you were promised

A sober counterfactual: imagine the dollar's value had been held roughly steady since 1980, and productivity gains had been allowed to flow through as falling prices.

Many staples would simply cost less in absolute dollars now than they did then, not more — a pound of bacon, a dozen eggs, a tank of gas. A starter home would be a smaller multiple of the median income rather than a larger one. A four-year college degree would cost what it cost your parents in nominal terms. Health insurance would consume a smaller share of the paycheck. Savings would compound automatically with the productivity curve — leaving a dollar in a checking account would buy more next year than this year, without doing anything, because the goods got cheaper while the dollar held its size.

That counterfactual isn't a fantasy. It's roughly what the second half of the 19th century in the United States actually looked like. The economy of 1900 was vastly larger and more productive than the economy of 1870, and the price level was lower. Real wages had risen. Workers got cheaper goods on top of higher pay.

Today's policy commitment is that this world is the wrong target — the 2% inflation goal effectively absorbs productivity into the dollar's decline rather than paying it back to dollar holders. Whether that's a defensible trade-off is a separate question. What's not a separate question is that the trade is happening, and you're on one side of it.

What this has to do with the clock

The clock isn't a forecast. It's the visible rate at which productivity-driven cost reductions show up in a denominator that can hold them — one where the supply is fixed and policy cannot expand it.

Bitcoin's supply runs on a hard, halving-driven schedule with a 21-million cap. As real-world productivity gains continue to compress the labor and resource cost of producing everyday goods, fewer satoshis are needed to buy the same real basket. That's what each item card on Satoshi's Clock is tracking. Not “Bitcoin going up” against the dollar. The productivity dividend showing up where the dollar wasn't allowed to absorb it.