The most expensive industry on earth is getting cheap

The most expensive industry on earth is getting cheap, and we should pay attention.

Defense costs have only ever gone one way. Norman Augustine, who ran Lockheed Martin, did the math back in the 80s and concluded that at the rate things were going, by 2054 the entire US defense budget would buy a single aircraft. It was half a joke, except that for forty years the curve held almost perfectly.

Then Ukraine happened, and the interesting number isn't the price of a drone, it's the cost per effect. A navy doesn't buy ships, it buys outcomes: a strait watched for a week, a target neutralized. For seventy years each outcome got more expensive because the only way to get it was a bigger platform with a crew on board. Take the crew out, make the system expendable, and the same outcome costs a hundred times less. The marginal utility of a dollar flips: the tenth exquisite ship adds little to what nine could do, the ten thousandth cheap system adds almost as much as the first. It also breaks the industry's business model. Much of defense still runs on cost plus contracts, a margin on top of costs, so for decades the most profitable move was to let costs rise. Inflation wasn't a side effect of the system, it was the product…

That isn't really a defense story though, it's a technology story. The cost of compute has dropped by roughly a thousand times in twenty years, solar has become the cheapest electricity humans have ever produced, and every real breakthrough ends up doing the same thing, which is more output with fewer people at a lower price. Technology is deflationary by nature, and it always has been.

The problem is what we've built on top of it. Every major central bank targets around 2% inflation, and that number is not zero by accident. Governments borrow against growth that hasn't happened yet, and mortgages, pensions and equity valuations all assume that tomorrow's economy will be bigger than today's, in nominal terms, forever (that’s the problem). So we have a technology that pushes prices down, sitting inside a system that breaks if prices go down.

The way we've reconciled the two so far is debt. Jeff Booth wrote a great book on this: we print and borrow to cancel out the deflation technology creates, and the gap between the two grows every year. Irving Fisher described what happens when that mechanism finally stops working, in 1933, a few years after losing his own fortune in the crash. The investor side tells the same story. Capital expects 5 to 11% a year, because that's what an inflationary economy has always paid, while the heavy industry that builds ships and ammunition earns low single digits on its assets. Nobody funds that gap voluntarily, which is why governments are starting to take equity in private industry. The state is becoming a shareholder precisely because private capital's return expectations were set in a world technology is dismantling.

What makes it harder to see is that the deflation isn't evenly spread. William Baumol pointed out in the 60s that whatever technology touches gets cheap, while whatever it doesn't touch (rent, healthcare, education…) keeps getting more expensive. That's why your phone costs nothing and your rent costs everything, and it's not a bug, it's the model working exactly as designed.

Keynes told us in 1930 that by 2030 we'd be working fifteen hours a week, because productivity would make the rest unnecessary. He was kind of right about the productivity and wrong about the hours, and the gap between the two is the whole question. Larry Summers calls the result secular stagnation, Robert Gordon thinks the big gains are already behind us, Jeremy Rifkin thinks we're heading towards a zero marginal cost economy, and they can't all be right.

I unfortunately don't have the answer. What I do know is that every institution we rely on was built for the old world, and I don't see anyone building for the one that's actually coming… Maybe blockchain ?