Ever since 2000, American’s views on the economy have been heavily shaped by two ideas:
1. There was clearly a big tech stock bubble in 2000.
2. There was clearly a big house price bubble in 2006.
It’s possible that there was some sort of a bubble in each case. But as of today, there is no strong evidence for either proposition. It is not true that tech stocks were clearly overvalued in 2000, and it’s not true that house prices were clearly overvalued in 2006.
The tech bubble theory was based on the assumption that stock prices had gotten so high that they could only be justified in the tech industry went on to dominate the US economy, and at least some tech stocks would have to become massive successes, in order to offset the inevitable attrition of tech start-ups that did not do well.
Today, it seems like the optimistic forecasts were basically correct. Not necessarily exactly correct, but close enough that one can no longer claim that the extreme optimism of 2000 was clearly unwarranted. Big American tech companies completely dominate our stock market, and indeed they dominate the global stock market in a way that was almost unimaginable in 2000. At least unimaginable to most individual humans; the “wisdom of crowds” somehow foresaw this future.
The same is true of housing. Real housing prices are now just 1% below their 2006 peak. So it’s no longer clearly true that houses were greatly overvalued in 2006. Instead, other theories are equally plausible. Perhaps NIMBYism combined with permanently lower interest rates does justify permanently higher real house prices in the 21st century. Perhaps the 2006 immigration crackdown and the subsequent tightening lending standards and the steep drop in NGDP during 2008-09 caused the slump, and house prices are now returning to their appropriate level. The new normal of the 21st century. Or maybe this is another bubble. Who knows? But as of today, we can see that real house prices were not clearly overvalued in 2006. It’s at least debatable.
The two theories that have been discredited played a major role in shaping the views of Americans, including economists. Now that these theories have been discredited, Americans should revise their view of what happened to the economy over the past 20 years. Will they do so? Probably not.
This is an invitation for bright young academics that are not blinded by popular prejudice to take a fresh look at the whole bubble theory, especially the way these ideas shaped our view of the business cycle.
PS. Ever since I began blogging more than 11 years ago, I claimed that bubble theories were wrong. But I’m not going to take credit for predicting the strong surge in tech stocks and house prices—I did not anticipate this.
READER COMMENTS
robc
Oct 23 2020 at 2:14pm
Alternate reading: Both are in another bubble.
If you draw a best fit line thru that housing graph in your article, it sure looks like a bubble in 2006 and a little bit high today, but nothing like 2006.
The stock market…ehhhh, I think growth stocks are a bit overvalued and value stocks undervalued right now. But I could be wrong. I have made (small) bets that way with my 401k, it is slightly overweighted towards value stocks. That isn’t paying off (yet). My slight overweight towards midcaps is, however.
Scott Sumner
Oct 23 2020 at 4:10pm
That’s possible, but even so the earlier “bubble” theories have been discredited, as the 2000 and 2006 prices are no longer “clearly” too high. It’s at least debatable.
AJ
Oct 23 2020 at 5:44pm
Can you share the evidence you think discredits bubble theories?
Scott Sumner
Oct 24 2020 at 1:18pm
I presented the evidence in this post. Bubble theories were based on a presumption that prices in 2000 and 2006 were obviously too high, they were obviously irrational. Today that doesn’t look obvious at all.
More specifically, the claim was that tech companies were not likely to be fabulously profitable in the future, and that real housing prices always tended to return to “normal” after a boom. Neither claim looks correct at the moment. Look at the “FAANGs”
If one says prices “might” have been too high, I of course wouldn’t argue. Market prices always “might” be wrong. But that’s not the bubble theory.
Bubble theory has gone from something that looked obviously true to most pundits, to a theory that’s unproven.
AMT
Oct 23 2020 at 2:50pm
Sure, according to the efficient market hypothesis there can’t really be bubbles. You claim that there was not “clearly” a housing bubble, because eventually prices rebounded. I would disagree that prices recovering in the long, or even medium term prove there was not a bubble. The issue is why prices are what they are at any given time: whether those prices are rational (not overoptimistic). If the price is simply based on estimates of future risk/reward, does a “bubble” retrospectively occur anytime we end up significantly on the negative side?
E.g. if a stock has a 90% chance of doubling in value, or 10% chance of losing 50% of its value (e.g. based on whether a prospective contract goes through or not), we should expect the price to rise simply based on those expectations, and then if the contract fails, to quickly drop. Was that a “bubble,” or just rational pricing based on the expected value? You must actually calculate the expected value to be able to say if it was a bubble or not.
I think this type of explanation could easily fit with the tech stock example, but I’m not so sure for the housing market. Given how stable the actual quantity supplied and demanded is in the housing market, should we expect large swings in prices? It seems to me that you could say there was “clearly” a bubble, it just wasn’t a huge bubble, which is why it didn’t take too long for prices to recover. And if you expect nominal prices to grow at something like 2% per year, wouldn’t house prices still be well below (about 30% below) our expectations if 2006 prices were not (at least to some degree) a bubble?
(Investopedia says the definition of a bubble is when the price “greatly” exceeds the assets intrinsic value. But what % is required for an asset to be greatly overvalued? Could the housing market have been a small bubble?)
https://www.investopedia.com/terms/b/bubble.asp
Scott Sumner
Oct 23 2020 at 4:12pm
I think you are missing the point. The bubble theory was based on the notion that there was no rational justification for those prices. But obviously there was. Those tech companies became phenomenally successful.
AMT
Oct 23 2020 at 6:22pm
Well, I think you are missing my point since I said I thought it could easily work as an explanation for tech stocks…and I think my explanation fit with you saying “If you bought Amazon plus 10 other losers, you did very well,” if you just adjust the numbers in my example. But how does it also work for housing?
But anyway, even if you say that it wasn’t a bubble years ago because look where we are now, don’t the huge fluctuations say something? If the earlier peak wasn’t a bubble, what was going on in the market after it crashed? Apparently if prices were not overvalued at the peak, why were they so massively undervalued shortly after? That seems like a massive question that you have to answer, if you say there was no bubble and markets were rational about their expectations for the future.
robc
Oct 24 2020 at 11:54am
That is a great point! There was either a long bubble pre-crash or a short bubble post-crash.
Scott Sumner
Oct 24 2020 at 1:22pm
I explained in my post the shocks that might have moved prices up and down.
But the bubble theory needs more than unexplained price movements, you need to know which price is wrong. No one can explain the stock crash of October 19, 1987, but no one knows in the pre-crash price was too high, or the post crash too low. So bubble theory is useless for the 1987 crash.
robc
Oct 26 2020 at 8:48am
The people who bought in 2002 got a better RoR than those that bought in 2000.
It looks like the return on the NASDAQ composite was 4.5% annually if you bought at the 2000 peak and 13.5% if you bought at the 2002 bottom. The answer is that both were wrong – it could have been a bubble in both directions. 4.5% is too low, considering the expected variance in tech stocks. The Sharpe Ratio on that has to be crap.
13.5% is probably too high for an 18 year period, I don’t think the Tech market risk justifies that, but maybe it does.
I think if you waited until the NASDAQ recovered to its 2000 level in 2014, the return is about 14.5% annually over last 6 years, which smells of another bubble possibly. But short term variance is much more likely so maybe not.
I personally think the growth stocks are too high right now and value stocks are too low. But not enough for me to avoid the growth stocks (neither was 2000). I dollar cost average, I guess my investment strategy assumes there are no bubbles, or at least they don’t really matter in the long run. Which is funny that I argue with Scott over this, as my revealed belief is more in line with him…I think I believe in bubbles yet don’t believe in my personal ability to see them and profit from them. But some people can. Maybe.
Max
Oct 23 2020 at 3:10pm
I don’t agree to Your idea that later high valued tech stocks show that in the 2000s no tech bubble burst. The tech stocks of 2000 have nothing to do with todays google, Amazon or Apple. So it might well have been a bubble of bad tech stocks with no value ideas.
Alan Goldhammer
Oct 23 2020 at 4:00pm
I think this is the correct reading. The big Internet companies of today were around in 2000 but all of them had earnings. There were a large number of companies that had minuscule or no earnings but whose share prices were totally not reflective of the underlying business. The same thing happened in the 1980s with biotech. At one point there were close to a 1000 biotech companies though most were small and privately held. Only a handful ended up bringing a product to market and of those even fewer are still independent. I remember a presentation in 1984 Paine-Webber analyst who made a valiant attempt to ascribe value to a biotech company with zero earnings.
I think Scott is wrong on this one. The housing bust is more complicated and was a result of over leveraging, NINJA loans, and structure debt vehicles that almost nobody understood (always fun to watch ‘The Big Short’ which should be on Scott’s list of great films).
Scott Sumner
Oct 23 2020 at 4:15pm
That’s not true. Amazon had no earnings back then. No one knew which company would be the next Amazon, so any promising company saw its stock price go up. If you bought Amazon plus 10 other losers, you did very well.
Alan Goldhammer
Oct 23 2020 at 6:07pm
Amazon did not go public until 1997. the did have significant cash flows in the early days that were all plowed back into the business. The same was not true of many of those companies that failed. I can only find consolidated financials going back to 2005 and am too lazy to search 10K filings for earlier years. Certainly by 2005 they were profitable and you could tell by the EBIDTA that it was solid. The question was always what the price premium you wanted to pay for the stock and it was always out of my value oriented approach (which was a big mistake).
Scott Sumner
Oct 24 2020 at 1:25pm
First you said they had no earnings. Then I said Amazon had no earnings, and you switched to cash flows. But Amazon was one of the most famous examples cited back then to claim there was a bubble. And lots of companies with no cash flow have gone on to be successful.
I am responding to the claim that NASDAQ was a bubble; I’m not defending every single stock price.
Philo
Oct 23 2020 at 9:05pm
The housing price graph you reproduce shows what looks like a mountain from, say, 2001 to 2011, with a peak in 2006. The “mountain” is obviously, indubitably there. But it is a “bubble” only if it was a *semi-irrational* mountain, where the rising slope was irrational, though the declining slope was rational; that, I take it, is the definition of the term ‘bubble’. If the decline, 2006-11, was very improbable based on the evidence available during 2001-06, the rise during 2001-06 was not irrational, and the mountain is not a bubble. And (though this seems less likely) even if the rise *was* irrational, based (let us say) on the wishful expectation that certain very improbably favorable conditions would obtain in 2006, the mountain might still fail to qualify as a bubble. For those wildly improbable conditions (or others just as good) might, just by luck, actually have come about in 2006; but the ensuing 2006-11 decline might in turn have been irrational, based, perhaps, on unreasonably pessimistic projections for conditions in 2011. Maybe the rational price line, 2006-11, was flat or gently rising, instead of the actual steep decline; if so, the actual mountain was not a bubble.
Bubble theorists tend to assume that price declines are rational, but they are quick to label price rises “irrational,” without ever producing any proof of irrationality. This looks like mere pessimistic bias.
Michael Sandifer
Oct 23 2020 at 9:21pm
I’m continuing to research this, but so far, when I check historical metro area annual NGDP growth it is typically pretty close to local real estate cap rates, which is consistent with the general concept that average rates of return on capital should equal NGDP growth, in monetary equilibrium, and/or in the long run. For stocks for example, since 1948, the difference in the average S&P 500 earnings yield and average NGDP growth is about 3 tenths of 1 percent. Also, the nature of the deviations between NGDP growth rates and the earnings yield are exactly what one would expect if they should be equal in monetary equilibrium.
Then, it’s easy to calculate an expected change in metro real estate prices, given the change in metro NGDP. So far, the numbers are all consistent with my hypothesis.
So far, I see no evidence for a bubble in US real estate, ever. Real estate appears to have been fairly priced, with what many would find to be surprising precision, all along.
The same is true of stock market bubbles. There’s no reason to believe they exist, and in fact, there are many good reasons to believe otherwise.
milljas
Oct 23 2020 at 11:06pm
And two years later?
Thomas Hutcheson
Oct 24 2020 at 8:16am
So what? Is it the only thing standing between us and a revenue neutral tax on CO2, a progressive consumption tax that gives us a small surplus at full employment, finance of retirement and health insurance with a AT instead of a wage tax, NGDP targeting, free trade, or a Matt-Yeglesian increase in H1B type visas?
bill
Oct 24 2020 at 7:38pm
I think every single person who was saying “this is a housing bubble” in 2006 had already been saying it since at least 2004, often sooner. Yet even the people that bought in 2006 did well. They got the expected dividend (a place to live) and appreciation equivalent to inflation. At any point in time, you’d expect to have a 50/50 chance of doing worse than that. So the bubble-sayers in 2002,, 2003 and 2004 probably talked many sensible people out of buying at a better than average point in time. We talk today about tulipmania because the prices are still down, on a real basis, after 350 years. Not because they took 14 years or more to recover.
Scott Sumner
Oct 25 2020 at 6:23pm
Yup. The Economist said housing was a bubble in 2003, and later bragged that they were right, even at a time when real housing prices were actually higher than in 2003!
Kurt Schuler
Oct 25 2020 at 12:12pm
Saying that a bubble did not occur 15 years ago because the real price today is just as high is like saying it was wise to erect a building 15 years ago that was empty until suddenly filling up last month, because the building is now full. The proper standard of comparison is the opportunity cost of a different allocation of resources. I think economists don’t yet have the right framework for thinking about bubbles, and that when they develop it, rather than starting from efficient markets it will start from something closer to Arnold Kling’s more relaxed view that “Markets fail–use markets.”
Scott Sumner
Oct 25 2020 at 6:21pm
Kurt, You said:
“Saying that a bubble did not occur 15 years ago because the real price today is just as high is like saying it was wise to erect a building 15 years ago that was empty until suddenly filling up last month, because the building is now full.”
That’s not the right analogy. A bubble claim is a claim that a price is CLEARLY wrong. Not just wrong, but CLEARLY wrong, ex ante. (All prices are somewhat wrong, ex post.) A building that lies empty for 15 years is very likely a bad investment, a waste of resources, whereas the fact that Amazon didn’t become profitable for ten years is not necessarily a sign it was overpriced at all, nor a waste of resources.
Today’s prices give us more and better information about the appropriateness of previous prices. In 10 more years we’ll know even more. Filling up a building that was empty for 15 years doesn’t eliminate the fact that it was built too soon.
My claim is that as of today, there is no reason to believe these markets were clearly wrong in their pricing in 2000 and 2006. There’s no evidence to support the bubble claim. You can argue that we don’t know for sure and that their might have been a bubble, and I’d agree. I’m merely saying that there currently no evidence to support the bubble hypothesis.
In the 1930s, it looked like the stock market was overpriced in 1929. Recent studies suggest that 1929 stock prices were reasonable. Ditto for mid-1987. As time goes by you get more and better information to evaluate whether previous prices were rational, although you never know for certain.
I do agree with one point you made. To the extent that “bubbles” are a problem of some sort, it would show up as a misallocation of resources. Of course misallocation can occur without price bubbles, and presumably some misallocation occurred in high tech. But yes, to the extent that bubbles are a problem it’s misallocation that is the actual concern.
Stéphane Couvreur
Oct 26 2020 at 5:43am
Yes, prices are always a little bit wrong ex post. So, I would add something to Kurt’s analogy: “Saying that a bubble did not occur 15 years ago because the real price today is just as high is like saying it was wise to erect a lot of buildings 15 years ago that were empty until suddenly filling up last month, because the building are now full.”
In other words, it’s not enough to have investment errors to make a bubbles, there must be a “cluster of errors”—an unusually large number of bad investments.
Best, Stéphane
Garrett
Oct 26 2020 at 8:55am
Tesla is another example. Guess where on this chart people were calling it a bubble. Trick question! The answer is everywhere.
Or how about bitcoin. As Scott has pointed out numerous times during the years, the bubblistas called victory at the beginning of 2018. But don’t forget that these same people had been calling it a bubble in 2017.
Doesn’t look so crazy on a log chart though. People who believe in bubbles don’t understand random walks.
Knut P. Heen
Oct 26 2020 at 12:21pm
Bubbles cannot occur in a market that allows either short-sales or production of new assets. During the so-called tech bubble there were many IPOs (production of new tech stocks/companies). Many of history’s biggest IPO underpricings happen during this period. The market supplied what the market demanded. If anyone argues that this was a bubble, they must claim that the IPO volume during the period was to scarce too keep prices down.
The housing bubble is more plausible. It is not possible to short housing. Zoning laws may prevent production of new houses. Location is also an issue. Yet, most of the people who says it was a bubble didn’t sell their own house. Why do they hold on to a clearly overpriced asset?
Todd Moodey
Oct 27 2020 at 10:44am
Unfortunately, for many reasons short-selling is woefully underproduced in U.S. financial markets. It’s even more true in Europe and Asia, where short-selling is routinely banned at the slightest hint of market weakness. Among these reasons are the awesome but indiscriminate transactional force of passive investing and a “Vanguardized” market, which has largely eliminated the discipline that short-selling and other beneficial market institutions used to place on markets. They are relatively new forces, and are likely to prove very problematic to the health of financial markets in the future.
Garrett
Oct 27 2020 at 2:52pm
If you could make money short-selling but are hampered by restrictions, then you could still outperform in a long-only mutual fund by not owning the stocks you’d like to short and owning the rest of the market. That active mutual fund managers still can’t seem to persistently outperform even with such an obvious opportunity leads me to believe that this isn’t a big deal.
Todd Moodey
Oct 27 2020 at 3:16pm
Among other things, there are non-zero costs associated with short selling that wouldn’t exist in your (non-held stocks = shorting) portfolio.
robc
Oct 27 2020 at 10:09am
John Paulson did it. As did Michael Burry.
They used credit default swaps instead of direct shorting.
Knut P. Heen
Oct 27 2020 at 12:14pm
You cannot sell a house you don’t own and thus increase the supply of houses. You have to build houses to increase supply.
You can, however, print a contract promising to pay the same dividend as Microsoft until date X. At date X, you promise pay the price of the Microsoft share. If you sell this contract, you have shorted Microsoft and increased the supply of Microsoft shares (except the right to vote at the annual meeting).
A credit default swap is simply insurance against default. You pay an insurance premium to AIG, for instance, and you get your money back from AIG if the issuer of the debt defaults. In exchange, AIG become the lucky owner of the debt contract in the case of default. Buying a credit default swap will not increase the supply of houses. It simply transfer losses from owners of sub-prime mortgages to AIG in the case of default.
robc
Oct 28 2020 at 8:58am
I would say your examples are the same thing. Your contract doesn’t create more MSFT stock just like the CDS doesn’t create houses.
The “important” part of house is living in it. The “important” part of the stock is voting rights.
Although my comment wasn’t exactly right either. Paulson et al didnt short housing, they shorted mortgages. Even with a “bubble”, the mortgages wouldn’t have failed at the same rate if the homeowners had had enough equity.
And naked CDSs arent insurance, as you don’t have to own the mortgage to buy one. If you own a mortgage and buy a CDS, you are taking out insurance. If you just buy one because you think the mortgage will fail, this is exactly the same as buying the MSFT contract because you think MSFT will pay a dividend.
Knut P. Heen
Oct 28 2020 at 1:22pm
No standard valuation technique put any weight on the value of the voting right (neither discounted cash flow nor the multiples-approach).
If you look at corporations with dual-class shares, the shares with the strongest voting rights trade at a premium. In the US the premium is usually less than 10 percent.
I would agree with you if the important value of a stock was its voting rights. Anyway, the point is that the cash flow from my contract and the Microsoft stock is identical, hence they are almost perfect substitutes (lacking the voting rights). Generally, financial economists think that all stocks and bonds are almost perfect substitutes because it is simply claims to future cash.
I agree that holding a naked CDS is not insurance. It is the opposite. It is pure gambling. It is similar to buying a put option on a stock you don’t own. You are betting it will go down. But again, buying derivatives does not affect the supply of the underlying asset directly (it may do indirectly if people learn that Paulson think these mortgages are junk).
In order for this argument to work on the housing market, the following chain of events must hold. Paulson buys naked CDS on mortgages. The market issues less mortgages because they learned Paulson was skeptical. Less supply of mortgages reduces bids for housing. This chain of events would potentially prevent a bubble in housing. I would not rule it out as impossible, but there are a few weak links in the argument.
By the way, I think the Tulip-bubble happened because it took 7-12 years to produce more of the specific type of Tulip they were “fighting” over. If someone wants to pay $1 billion for an asset that cannot be multiplied by production, that is the market price. If the guy sells it for $1 million the day after, that is the new market price. We may call it a bubble, but there is nothing anyone can do to stop it from happening.
Comments are closed.