Some Bubbles Are Easy: Five Lessons from Ben Inker

GMO Head of Asset Allocation on earnings bubbles, six railway lines to Manchester, and why surviving a bubble is harder than spotting one.

Ben Inker has been at GMO long enough to have lived through four bubbles, by his own count, in twenty-six years. So when he sat down with us for the first time since 2021 and described the current AI boom as an “easy bubble,” the phrase was clearly doing real work. Easy, in Ben’s usage, refers to the escape routes rather than the ending: can an investor sidestep the danger while still holding a portfolio that makes sense if the danger never arrives? Some bubbles permit that; the worst ones force a choice between prudence and keeping your clients. Over the hour we worked through GMO’s recent research with him: the case that today’s excess may live in earnings rather than valuations, the long history of world-changing capital spending, the origin story of the firm’s benchmark-free portfolio, and the question Ben kept returning to: why, exactly, are you getting paid?

Lesson 1: Judge a Bubble by How Hard It Is to Escape

A bubble, as Ben’s team defines one for portfolio purposes, is a situation where some important asset has risen to a level where it feels “pretty close to unownable.” The definition matters less than the question that follows from it. If you believe you are in a bubble, how hard is it to build a portfolio that avoids the worst of the pain and still lets you keep your clients? The internet bubble, on this scale, was easy. The S&P 500 was the most expensive it had ever been, but an investor could still hold a normal sixty percent in stocks by swapping large cap growth for small caps, REITs, and emerging market equity and debt. If the bubble call proved wrong and the world turned out normal, that portfolio still earned a normal return. 2007 was hard. Every risk asset GMO could find looked massively overvalued, so escape meant abandoning risk itself. Late 2021 was harder still, because stocks and bonds were overvalued together; the only refuge was cash, and cash is the one asset no client will tolerate paying for.

You do not need a professional money manager to own cash. You can do that in the bank.

Ben was candid about the cost of testing that leash. A manager who runs a portfolio that will look very stupid unless the world falls apart quickly gets fired, and the client overwhelmingly hires “the person who was doing the exact opposite of you.” That is how clients in 2000 made less money on the way up and then lost more on the way down. Today’s bubble, he argued, is back in the easy column: US stocks look overvalued, while risk assets outside the US are still priced to deliver a decent return.

Lesson 2: Sometimes the Earnings Are the Bubble

We put the standard objection to him. In 2000, the giant stocks traded at bubble valuations; today, setting aside names like Tesla, Palantir, and SpaceX, a company like Microsoft trades at multiples that look expensive rather than crazy. Ben granted all of it, then moved the problem somewhere most investors never think to look.

A tricky thing about today is we worry that this may be an earnings bubble. 2000 was a little bit of that, but mostly a valuation bubble.

Europe showed what one looks like. Earnings there rose 100 percent in the four years before 2008, and on an index basis they are still struggling to climb back to those levels. Today the mechanism runs through capital spending. When Microsoft spends 200 billion dollars on data centers, none of that spending comes out of Microsoft’s income, yet all of it lands as somebody else’s revenue. Depreciation eventually claws it back, but slowly, and a data center that takes three years to build has not started depreciating at all. The current wave of investment flatters profits at both ends; Ben called this a very good time for corporate earnings, and probably an unsustainably good one.

Anyone who has owned resource stocks knows the endgame. You do not buy the miners when their P/Es are low, because a low multiple means peak earnings and, as Ben put it, “the capital cycle is about to come around and bite you on the ass.” You buy when the multiple is high and the earnings have collapsed. Applied to the AI supply chain, that logic yields a specific forecast: before this cycle ends, the SK Hynixes and Microns of the world will look really cheap on trailing earnings and still turn out to be lousy investments.

Lesson 3: Don’t Assume the Profits Go to the Builders

Start with scale. Data center spending this year is forecast at something like 700 billion dollars, around 2.2 percent of US GDP. That makes it a bit bigger than the fiber optic build of the late 1990s, about half the size, relative to the economy, of the railroad spend in the second half of the 1800s. Ben’s reading of those precedents begins with a concession to the optimists: the people who believed canals, railroads, electricity, automobiles, and the internet would change the world were right every single time.

Just because something changes the world doesn’t necessarily mean the profits accrue to the people who built it.

Railroads carried his argument. They collapsed the cost of long-distance transportation, which made specialized manufacturing and agriculture possible; California, he observed, could not really have existed before them. Running a railroad, though, was never an amazing return on capital, and investors kept making it worse. The first company to lay track between London and Manchester would have made a lot of money had the building stopped there. Five more companies followed, and with six railway lines running between the same two cities, the overcapacity destroyed the return on all of that capital, the early dollars along with the marginal ones. Fiber repeated the pattern, a lousy return that enabled the world we live in now. Unless something outside the market interrupts the spending, and a shortage of electricity might, Ben expects capital to keep flowing until the returns are gone. Should demand somehow stay ahead of supply indefinitely, he allowed, “that would truly be a this time is different situation.”

Lesson 4: Never Own Something Out of Fear It Might Go Up

In the fall of 1999, GMO was running benchmarked sixty-forty portfolios, underperforming, and fielding complaints, which is what underperformance produces. What struck Ben was that the complaints contradicted each other. One set of clients looked at a benchmark that was fifty percent S&P 500, saw GMO holding just twenty-five points in US large caps, and demanded to know how it could make sense to spend that much of a risk budget on one bet. Their verdict, in Ben’s retelling: “That’s stupid.” A second set of clients looked at the same portfolio and asked the opposite question. GMO was forecasting a negative real return for US large caps over the next ten years, so why was a quarter of the portfolio sitting in a risky asset expected to lose money?

GMO was running one portfolio for clients thinking about it in two incompatible ways, and the second complaint had no good answer. Why did GMO own twenty-five points of an asset class it hated?

It was out of fear that they might do well, despite the fact that we hated them.

That fall the firm started showing clients what the portfolio would look like with no benchmark at all. It was 1999, and nobody was interested; even the clients who had not yet fired GMO were not asking for more of what it was doing. Then 1999 turned into 2000, 2000 turned into 2001, and the first clients began to concede the idea made some sense. The strategy launched in 2001 with a single promise: everything in the portfolio has to make sense on its own, and nothing gets held out of fear. The useful question for the rest of us is which risk we actually care about, absolute or relative; a portfolio built to soothe both anxieties at once will do neither job well.

Lesson 5: Ask Who’s Taking the Other Side of Your Trade

We closed by asking Ben how to apply the idea that runs underneath nearly all of his work, understanding why you are getting paid for what you own. He began by shrinking the claim. The question will not make you rich; “the way you get rich, the way you massively outperform, is by predicting the future better than somebody else or everybody else, or just getting lucky.” Its real use is protective. Consider the pitch for a tail risk hedge that earns a cash-like return, protection in bad times at no long-run cost. Ben’s test is to stand on the other side of the trade. Somebody has to agree to lose a ton of money precisely when the world is falling apart, in exchange for cash-like compensation, over and over. No one can sensibly keep doing that, and the vast majority of people who bought such products, he noted, wound up really disappointed. A permanent diet of call options fails the same exam; the buyer pays enough each time, in Ben’s words, “to suck out all of the goodness of the stock market.”

Then he turned the same weapon on his own tribe. Junky, levered companies are more likely to go bust in bad times, so in theory owning them should pay; in practice they have been a horrible group to own, and that claim survives only as long as the pricing does, because “any horrible group of stocks, if priced cheaply enough, becomes a wonderful group of stocks.” Ben loves value conceptually. That did not soften what came next.

There is nothing more supremely useless than overvalued value stocks.

They have, he added, no virtue in your portfolio whatsoever.

The Bottom Line: Avoiding Stupid Is Most of the Job

GMO’s seven-year forecasts rest on a belief he stated plainly: “It would be weird if we lived in a world where the return on capital and the cost of capital did not come into alignment with each other. That’s what capitalism is supposed to do.” Every failure we talked about is a way of getting hurt while waiting for that alignment, whether by owning assets out of fear they might go up, trusting earnings a capital spending boom has temporarily inflated, assuming the people building the future must end up its beneficiaries, or buying protection nobody sane would keep selling. Nothing in Ben’s toolkit promises the timing, which is why even an easy bubble is only easy for the investor who does not need it to burst on schedule. He priced the discipline honestly, and modestly. “It isn’t the secret to success, but I do think it can keep you out of a lot of painful kinds of failure.”

Watch the full episode here: