Between the Lines

What to Do When You’ve Spotted a Bubble

Yes, you could short it — but what if you’re a few years early?

Credit: Kaveh Kazemi/Getty

Suppose you’ve found an asset you believe is overvalued. Maybe it’s a currency, a bond, a commodity, a stock, a sector, whatever.

What, exactly, should you do with this information?

Let’s assume you’re smart. Your thesis is correct, it’s overpriced, some day the price will correct, and lots of people will get hurt. But for the asset to be overpriced, somebody has to disagree with you.

So what do you do when you spot a bubble? This is an incredibly broad question, but it’s a question all investors in public markets have to ask. There are three basic schools of thought:

The brave, stupid, and intellectually consistent approach

You could just short it.

Sometimes, that works.

Sometimes it doesn’t, though. During the housing bubble, one manager wrote up a very compelling case for why subprime debt would end badly. The subtitle of it was “A Home Without Equity Is Just a Rental With Debt.” Anyone armed with that information would have been unsurprised by what happened from 2007 through 2009.

Here’s the problem: the title right above that subtitle was “Housing In the New Millennium,” because the write-up was from 2001. If you’d read the essay then, you might have done the obvious thing and shorted a homebuilder like Toll Brothers, which, at the time, traded at about $9 per share. TOL peaked in mid-2005 at about $57 per share.

Other homebuilders had similar performance. Bank stocks also took a long time to break down.

Of course, prices did eventually collapse, but shorting a basket of homebuilding stocks would have been unsatisfactory to an average investor and a fireable offense to a professional.

Normally, the quote people cite in situations like this is from Keynes, who apocryphally said, “Markets can remain irrational longer than you can remain solvent.” But a more apt quote is, “A bull market climbs a wall of worry.” The risk of a housing downturn is the risk a homebuilding investor is paid to take. That’s why they nearly always trade at single-digit, price-earnings ratios.

Of course, some short-sellers do put up enviable track records, and some people have famously shorted bubbles at exactly the right time.

Merely knowing there’s a bubble is a good way to get run over.

It’s hard, though, because typically bubbly assets are priced such that they get higher returns than comparable assets, just returns not quite high enough to compensate for the risk. Subprime debt did yield more than other kinds of paper, just not enough more.

Bubbles also tend to persist for a long time, despite their intrinsic instability because they’re partly self-perpetuating. When subprime mortgages drive real estate demand and higher real estate prices let bad borrowers periodically refinance, the debt actually has a lower default rate than lenders expected. When small dot-com companies invest their IPO dollars in display ads for big companies, and big companies invest their display ad revenues in servers and software licenses, IPO proceeds turn into revenue, revenue turns into market cap, and market cap justifies the next round of IPOs.

Merely knowing there’s a bubble is a good way to get run over.

The mercenary, but profitable approach

A fair warning: This approach is most popular among people who get paid an annual bonus but invest in assets that have a bad year every five or 10 years.

In the book Hedge Fund Market Wizards by Jack Schwager, there’s an interview with a macro trader. The trader says it’s crazy to short bubbles. (Correct!) He concedes that during a bubble, the true believers will have higher returns (He’s on a roll!). So, what should you do during a bubble instead? What did he do in 2006 and 2007?

“What I believe in is compounding and not losing money. We were quite happy to be part of the bubble, but to do it in positions that were highly liquid, so that we could exit the market quickly if we wanted to.”

This is not a bad approach. It’s intellectually inconsistent, but that’s okay; everyone needs a little intellectual inconsistency to get by.

Boiling it down, the argument is if you’ve correctly mapped the wall of worry, you have a competitive advantage when it comes to worrying. You buy your ticket, take your seat, and watch the show, but when you see a few people get up and head for the exits, taking great pains not to walk too fast, you join them.

This is a scary way to live. It’s like having an affair or being a professional drug dealer. Probably a thrill, sure, but at any moment you know you’ll need to make some very difficult decisions, fast.

You have to make them especially fast because many, many asset managers pursue the strategy of riding bubbles just long enough and then are first to sell. Since it’s dangerous to do this with just one, they’ll pick a few at once, hoping they won’t all collapse at the same time. This diversification allows them to lever up a bit, which juices returns.

If some of your portfolio is invested in high-flying momentum stocks, some is in leveraged loans, and some is in currency carry trades, you are more diversified than someone who does just one of the above.

The math of diversifying away risks is brutal, though: When your bets start to correlate, your implied risk explodes. Suppose you have three uncorrelated bets that each have a 10% chance of dropping 10% in a given year. You’d expect a 10% loss once a decade, a 20% hit every century, and 30% every millennium. So unless you’re running the Long Now Foundation’s endowment, you can feel comfortably levered over 4-to-1. Now, suppose all those bets are 100% correlated: Your expected loss in a bad year triples. Now, the safe level of leverage is at most a bit over 3-to-1, but perhaps you’re correctly nervous about how volatile your risk models are, so you go below that.

Risk managers take a prudent approach. As measured risk rises, they want to take leverage down. But taking leverage down means unwinding several trades at once — selling your wild SaaS stocks, dumping your leveraged loans, selling off some pesos and paying back some yen. And all that activity both hurts the returns of traders who followed your strategy and raises correlations, forcing them to sell more, too. If you ever want a really quick way to see if this is happening, look for a day when the only stocks that are up are the ones with high short interest.

I idly wondered what happened to that Hedge Fund Market Wizards trader after the interview. Turns out he was short the Swiss Franc when it got revalued by 40% overnight. He shut down the fund and runs his own money now.

Hence my warning at the beginning. The annual bonus is key.

Nirvana: positive carry and positive skew

Hedge Fund Market Wizards is a good book to read because it’s part of a mini-genre of books explaining why Michael Lewis was wrong. If you read The Big Short and wondered how a tiny fund like Cornwall Capital—started with $100,000—could swing such big bets just a few years later, wonder no more: Lewis omitted some details that make the story less exciting but more plausible.

I’m going to humbly submit my own contribution to the Lewis-is-wrong field: The Big Short was actually not a great trade. If he were really on the ball, he would have written The Big Pair Trade.

Timing bubbles is hard. Most of the people profiled in The Big Short were extremely smart, and many of them did, in fact, do enough homework to time the bubble well. But it wasn’t a unique thesis. Google Trends says that searches for the term “housing bubble” peaked in August 2005. The idea was already in the air. In 2005, you could have talked to a lot of people who wanted to short housing, most of whom ultimately lost money.

But some more homework would reveal an obscure trade that correctly identified the pathologies of high housing prices and subprime debt and actually offered positive returns during the bubble.

To understand why, we have to delve into the details of subprime collateralized debt obligations.

Slicing the salami: How to make a collateralized debt obligation

You can describe CDOs by analogy. They’re like a bond fund whose investors own debt rather than equity. Or they’re a savings-and-loan in an alternative universe in which Wylie Dufresne has the same personality but works at an investment bank.

A CDO is salami — you start with ingredients that are edible but unappetizing.

All of these are misleading. A CDO is salami — you start with ingredients that are edible but unappetizing. You blend them together, process them a bit, add a little spice, slice, and serve.

Specifically, you take a package of debts that have less than 100% odds of repayment, and you offer investors one tranche, typically the largest, they get paid unless a staggering number of the debts default. If the average default rate is 10%, you might see these investors getting paid so long as 30% of the bonds are not in default.

Just by the logic of diversification, this debt is more creditworthy than the constituents of the overall debt package; the first defaults cost it zero.

To equalize things, you have other tranches, which take on the next layer of default risk and have correspondingly higher yields. Finally, you top off the whole structure with some collateral: A small sliver of extra equity that takes the first loss from defaults. (Otherwise, the highest-interest and lowest-rated slice of the CDO would be the junkiest of junk credits.)

All this is a simple consequence of a fundamental financial fact: Assuming prices are accurate, diversification is a free lunch. A given asset has an expected return, some overall volatility, and some volatility idiosyncratic to it. Pick two, and some of the idiosyncrasies cancel out. Pick two hundred, and almost all the idiosyncratic volatility disappears, leading to higher risk-adjusted returns.

That assumes you can measure the volatility, and that it is indeed idiosyncratic to each security.

The big pair trade

Returning to that self-perpetuating mortgage bubble: If housing price appreciation slows defaults, and housing prices are a function of subprime debt availability, then to the extent that subprime loans end up in the same CDO, they’re more correlated than history suggests.

This has two effects. As most of the protagonists of The Big Short figured out, it means AAA-rated slices of subprime-backed CDOs were not, in fact, worthy of the AAA rating. But it also means that the riskiest portion, the equity tranche, is less risky than it looks. High correlation within a pool of securities means mass defaults are more likely, but it also means that zero defaults are more likely.

Since the equity tranche offered exceptionally high returns in the event that defaults turned out to be low, an investor could buy a small slice of the equity, buy insurance against a big chunk of the AAA-rated debt, and end up with a trade that had the following characteristics:

  1. The equity’s returns more than compensated for the cost of buying insurance against the AAA-rated slice.
  2. It would make a little money if real estate drifted up, because the equity would be worth more while the AAA insurance couldn’t be worth much less.
  3. It would make a lot of money if real estate dropped.
  4. High correlations within the same bundle of mortgages would imply that the idiosyncratic returns were at the CDO level, rather than at the mortgage level, so it would benefit from high diversification.

A few people figured this out. I dearly wish I had numbers on how much they made. (We have hints because some of them did so well they got sued.) Most of the media reports speculated that the people doing this trade were buying the equity slice to fund deals they knew would blow up, but that doesn’t really make sense. Banks used to keep the equity because even though it would probably go to zero, they’d be more than compensated by the fees on the rest of the transaction.

This whole trade drops the adversarial nature of shorting a bubble and the sellout nature of rolling with it. Instead of saying “You’re dead wrong,” or “You’re absolutely right,” it offers a subtler critique: “If you think that’s true, wouldn’t that imply…” Since bubbles have internal logic, and incorrect internal logic is the essence of the bubble, this process means finding cases where the contradictions in the logic are expressed in the prices of two different tradeable assets.

Some bubbles offer opportunities like that; if you look hard, for example, you can find the same kinds of positive-carry/positive-skew situations in betting on the rise of passive investing. Some bubbles don’t, or if they do it’s quite subtle. I don’t know of any part of the internet bubble that was amenable to this, for example.

All this is actually pretty fair. There is no free lunch after all. You can virtuously lose money tilting at windmills by shorting a bubble. You can make money (until you blow up spectacularly) by tagging along. Or you can arbitrage the internal contradictions of the bubble. But to do that, you have to be absolutely sure you’re right, and you have to work hard to find a way to express the trade.

I write about technology (more logos than techne) and economics. Newsletter:

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store