In one of the defining sequences of The Big Short, a film about the 2007-08 financial crisis, a team of hedge fund managers go to Florida. They meet a pair of sleazy mortgage brokers who reveal how the US housing market is propped up by subprime loans to strippers, people with no job or income and immigrants who do not understand the paperwork. One of the managers calls up his trader and declares: “Hey, there’s a bubble.”
The film, based on the Michael Lewis book of the same name, goes on to track how a small group of investors make billions by betting against the market as the bubble bursts, causing disaster across the global financial system and triggering one of the worst recessions in modern history.
It lays out what has become the standard story of the crisis: new financial instruments allowed speculators and low-income households to go on a wild borrowing binge, inflating an almighty boom in the US housing market, which then collapsed to devastating effect.
It is a powerful tale. But is it true? According to new research by economists including Gabriel Chodorow-Reich of Harvard University, it is not necessarily a full description of what happened. Looking back on 2007-08 with what they call “2020 hindsight”, Chodorow-Reich and his colleagues suggest that while speculative excess was certainly important, at least some of the US housing market’s rise was rooted in economic fundamentals.
Their main line of evidence is rather simple: US house prices have more than bounced back from their 2007-08 crash to reach new highs. Where the Case-Shiller US national house price index peaked in July 2006 at 184.6, it stands today at 260.9.
“The 2000s housing cycle was not a boom-bust,” says Chodorow-Reich, “but rather a boom-bust-rebound”. If the price levels of 2007 were speculative and irrational, but housing markets quickly regain and surpass them, it raises a question about just how irrational the prices in 2007 really were.
This pattern of boom-bust-rebound is repeated in some other famous bubbles. As the Nobel-prize winning economist Robert Shiller once noted, history does not generally support the idea of bubbles that catastrophically and irrevocably burst. “Though the abrupt ends of stock market booms in 1929, 2000 and 2007 might seem consonant with such a metaphor, these booms were reflated again before long, [in] 1933-37, 2003-2007, and 2009-present respectively,” he said.
Today, the shares of US technology companies, which have reshaped the global economy in the past three decades, trade at lofty new peaks. Apple is valued at $2.6tn; Microsoft at $2.7tn. Relatively young companies have reached eye-watering valuations — Tesla trades at 20 times sales — and online traders have driven the price of so-called “meme stocks”, such as video game retailer Gamestop, to the moon and back again. The gains have encouraged investors everywhere, including in the UK, to put more money into tech.
They have also prompted some prominent commentators to express fears of a new bubble. “It seems to me very clear that the stock market is extremely overvalued and this is very dangerous,” says Andrew Smithers, an economist who has written extensively on how to value the stock market. “Bubbles usually end in tears.”
So what is an FT Money reader to make of it all? Investors want to protect their savings from a crash, but they also want to protect them from the guaranteed loss of purchasing power involved in staying out of the stock market at a time when interest rates are lower than inflation.
It is notoriously hard to invest during a bubble — even if you diagnose one correctly, the wait for it to burst can be unendurable — but in today’s stock markets, it is worth trying to understand more about the nature of bubbles and the relevance of a subsequent rebound in prices.
After the bust, the rebound
The simple fact that US house prices bounced back after 2009 says little about whether there was a bubble in 2006-07. It could be that something changed after 2009 to make houses more valuable, such as the fall in global interest rates, or house prices in 2021 could also be in a bubble. It is never possible to infer much from a simple change in price — but the evidence compiled by Chodorow-Reich and his co-authors, Adam Guren and Timothy McQuade, is subtler than that.
“It’s not just a boom-bust-rebound in national data but also in local data,” says Chodorow-Reich. “Those places that had the biggest booms had the biggest busts and the biggest rebounds.” There is a strong correlation between an area’s house price growth in the “bubble” period of 1997-2006 and its overall growth from 1997-2019.
Suppose that house prices in certain cities, or all cities, were overinflated by an irrational, speculative bubble in 2006. It makes sense that prices would crash the most in places where they were most overinflated. But if the original rise was irrational, why would they rebound in exactly the same places as before? It is like rolling a set of dice. If the same numbers keep coming up, you will wonder if the dice are truly random, or if some other force is at work.
Chodorow-Reich and his colleagues go on to show that the boom-bust-rebound in local areas is correlated with a set of supply and demand factors, such as land availability, zoning rules, wages, weather and even the number of restaurants. The implication is that these fundamentals have at least something to do with the rise in US house prices from 1997 to 2006 and onwards to today.
Fundamentals can explain a rise in house prices but they cannot, by themselves, explain a boom that gets ahead of itself, a crash and a subsequent rebound.
“To do that you need something else,” says Chodorow-Reich. “One is over-optimism in the boom. It depends how you define a bubble but we’d certainly agree that house prices in 2006 were above their long-run trend.” The other ingredient is foreclosures, so that the boom turns to bust. “The key thing about the housing market is everybody is levered, everybody has a mortgage,” he says.
The story then runs something like this. Around the middle of the 1990s, something happened to make the fundamentals of certain US cities more appealing. “Some of what changes is the industries that are taking off. It’s around the time that tech is becoming very important,” says Chodorow-Reich. “The start of the national boom is also coincident with the decline in crime in the US, which made central cities much more attractive again.”
Increased demand for housing runs up against a lack of supply, either because of geographical constraints such as water and mountains, or regulations that make it hard to build. That causes prices to rise. People notice that house prices are going up and they extrapolate, becoming over-optimistic and driving the market to an unsustainable high. Then at some point the good news runs out. Prices drop. Since the houses were bought with mortgages, that creates a downward spiral of foreclosures and falling prices.
Where this differs from the usual story is that neither subprime credit nor speculative investors play a vital role — which is not to say that they were not present. “Both things can be going on and they matter differently in different places,” says Chodorow-Reich. “One of the places that was a poster child for speculation was Las Vegas, where prices suddenly zoom up in the mid-2000s, with a high investor share.”
A bubble or not?
Does this mean there was no bubble in the US housing market in 2007-08? There was, after all, a huge boom and then a collapse in prices. The people who bet against the market made a great deal of money — there really was a Big Short — while investment banks such as Lehman Brothers went bust.
Ben Inker, head of asset allocation at GMO, a fund manager that has called several bubbles in the past, is not in such a hurry to rewrite the history of 2007-08. “The thing that I find difficult in reappraising and saying, actually house prices were fine, is that there were a bunch of us saying, ‘this is crazy, this cannot persist’, and we were right and the financial system took a big hit,” he says.
Chodorow-Reich and colleagues are quick to point out that this is not quite what they mean. Rather than asserting there was no housing bubble, their claim is that it stemmed from a real change in the fundamentals for housing, which is why the bust was followed by a rebound, rather than evaporating as if it were never there.
It comes down, in part, to the nature of a bubble and how it is defined. If speculation and irrationality are the essence of a bubble then Chodorow-Reich’s work calls that into question in 2007-08; but if a bubble is simply a period when prices are well above fair value, then there is no dispute. “I think that’s what a bubble is,” says Smithers. “A bubble is an extreme overvaluation.”
The crucial idea from Chodorow-Reich’s version of the 2007 housing boom is that it starts from an underlying change for the better. That is what allows the cycle of over-optimism to get started. A similar analysis could be made of technology stocks in 2000. Valuations then were about as extreme as the stock market has ever seen, but it is hard, in retrospect, to say that investors were wrong about the importance of the internet or the profits to be earned from moving business online. They were simply too early and had the wrong set of companies.
Yahoo’s market capitalisation peaked in January 2000 at around $125bn, at which point it was the most valuable company in the world. Today, Alphabet — owner of Google, which is in essentially the same search engine business — is capitalised at $1.92tn. Pets.com, an online retailer of pet food and accessories that became one of the symbols of the dotcom bust, reached a peak market capitalisation of $290m. Today, Chewy Inc — an online retailer of pet food and accessories — is capitalised at $32bn.
One question today is whether it is possible to detect a shift in fundamentals that could provide the impetus for a round of price rises and over-optimism. The pandemic has certainly not been a period of general optimism about the future. The best candidate is therefore the long-term slide in interest rates — which makes any given stream of future cash flows more valuable. This is combined with the rapid rise of online business models that require little capital investment, seem largely invulnerable to erosion by competitive pressures, and so look very valuable for investors.
“Prices in the US equity market look extreme relative to history, but they look less extreme relative to interest rates,” says Inker. “If this turns out not to be a bubble, the answer will be that the underlying environment for financial valuations has changed, so investing in stocks with a much lower expected return makes sense.”
“I’m sympathetic to the idea that with very low interest rates you can sustain higher equity valuations,” he says. “But it’s still very explicitly a bet on where real interest rates are going to be or that they’re going to continue to go lower.”
Shiller, who was forthright about the dotcom and housing bubbles, has also been hedging his bets this time around because of the low level of interest rates. “The stock market is already quite expensive,” he wrote in March. “But it is also true that stock prices are fairly reasonable right now.” The difference is whether the comparison is made with historic prices or with the relative yield available on bonds.
Only Smithers is ready to state clearly that interest rates do not matter and thus the market is overvalued. “You can demonstrate overvaluation without reference to the level of interest rates,” he says, pointing to two different yardsticks. One is the cyclically-adjusted price/earnings ratio, a measure of stock market valuation developed by Shiller, in which earnings are averaged over ten years to smooth out the economic cycle. The other is Tobin’s Q, a measure that compares the market value of corporate assets with what it would cost to replace them. Both are flirting with the levels recorded in 1929 and 2000.
The view that bubbles begin from some fundamental positive change is ultimately a frustrating one for investors. It means that there will always be a good story to tell about why prices are going up — and that story will be correct — but it can simultaneously be correct that prices have gone too far and entered a bubble. Chodorow-Reich suggests that is true of US housing 15 years ago; it could be true of technology stocks today.
That there is a debate even now about what happened in 2007-08 shows how hard a bubble is to diagnose, even after it has burst. To spot one in advance — that is the toughest job in financial markets.