Scott Sumner  

Optimal mutual funds in a world of bubbles

My Review of Barbara Bergmann,... Mankiw on Free Trade in the

One argument in favor of the Efficient Markets Hypothesis is that if bubbles existed, it should be possible to make large excess returns by betting on an eventual collapse of the bubble. And the famous counterargument often made is that:

The markets can stay irrational for longer than you can stay solvent.
Thus people who felt that tech stocks were overvalued in 1996, or American real estate was overvalued in 2003, and who shorted tech stocks or MBSs, might go bankrupt before their accurate predictions were finally vindicated.

There are lots of problems with this argument. First of all, it's not clear that stocks were overvalued in 1996, or that real estate was overvalued in 2003. Lots of people who made those claims later claimed that subsequent events had proven them correct, but it's not obvious why they were justified in making this claim. If you claim X is overvalued at time t, is it vindication if X later rises much higher, and then falls back to the levels of time t?

But let's put that common cognitive bias aside, and consider what the world would look like if bubbles were a real phenomenon. The first thing to note is that the term 'bubble' implies asset mis-pricing that is easily observable. A positive bubble is when asset prices are clearly irrationally high, and a negative bubble is when asset price are clearly irrationally low. If these bubbles existed, then investors could earn excess returns in a highly diversified contra-bubble fund. At any given time there are many assets that pundits think are overpriced, and many others that are seen as underpriced. These asset classes include stocks, bonds, foreign exchange, REITs, commodities, etc. And even within stocks there are many different sectors, biotech might be booming while oil is plunging. And then you have dozens of markets around the world that respond to local factors. So if you think QE has led Japanese equity prices to be overvalued, and tight money has led Swiss stocks to be undervalued, the fund could take appropriate short positions in Japanese stocks and long positions in Swiss stocks.

A highly diversified mutual fund that takes advantage of bubble mis-pricing should clearly outperform other investments, such as index funds. Or at least it should if the EMH is not true. I happen to think the EMH is true, or at least roughly true, and hence I don't actually expect to see the average contra-bubble fund do well. (Of course individual funds may do better or worse than average.)

The Economist recently reported that Warren Buffett has not done as well in recent years:

Mr Buffett used to argue that Berkshire's book value per share, rather than its share price, was a good proxy for its long-term worth. But the group's book value has stopped outperforming the broader stockmarket--in fact it has underperformed it in five of the past six years (see chart). So now Mr Buffett has begun to argue that book value is no longer such a good measure, and to give greater prominence to Berkshire's share price. This sort of goalpost-moving is a habit of lesser conglomerates than Berkshire, and is hardly a promising sign.

Comments and Sharing

COMMENTS (22 to date)
D. F. Linton writes:

If the recent years are the up ramp of a bubble, the fact that a value investor like Buffett hasn't performed well during this period hardly says anything about his performance over a period that include both the inflation and bursting of a bubble.

As John Hussman says in a bubble you only have two choices: Look like an idiot before it bursts or look like an idiot afterwards.

But maybe there are no bubbles and both of them are just wrong.

E. Harding writes:

The big problem is that recognizing the existence of a bubble is very different from having a good understanding of the timing of the bubble's popping. I don't think you've rebutted the "famous counterargument" here well. The whole counterargument depends on the crucial issue of timing the bubble, which is much more difficult from simply recognizing it.

A writes:

The Economist article isn't obviously evidence of EMH. You would not expect GAAP values to match the volatility of the S&P from 2009 - 2015. In fact, if a manager wanted to exploit the market pricing relative to cash assets, he/she might increase the payout ratio in 2009, thereby assuring book value underperformance relative to the broader market.

Maybe it's the case that the 13-Fs show underperformance.

One question about bubbles: Does the notion of a bubble imply reduced valuations elsewhere? Is there some reference, like NGDP, where we can say "For a given ___, a bubble here implies under valuation elsewhere"?

Scott Sumner writes:

DF, Mr. Buffett also lost a lot of money in 2008, so it isn't just the boom years.

E. Harding, I think you missed the point. The strategy I proposed does not require that you be able to predict the timing of bubbles, just their existence.

A, I don't believe that bubbles exist, so I can't answer your question. Maybe someone else can.

Lorenzo from Oz writes:

If we could get folk to ask the question "how stable are the prices of X asset?", it might improve discussion somewhat.

Australian households tend to be highly leveraged (via housing mortgages) on bureaucratic approval (via endemic land rationing), but the so far endless Australian housing "bubbles" are based on pretty stable factors which continue to be priced in.

Rajat writes:

Basically agree Scott. The only one that bothers me is the tech boom of the late 90s. Prices of companies with no clear business plan rising exponentially for no discernible reason. We had a similar thing here in Australia during the mid-70s resources boom. If the bubble concept is to mean anything, it must refer to a phenomenon that is rare, so maybe it wouldn't be possible to develop a contra-bubble fund.

Kenneth Duda writes:

Scott, I think it's worth pointing people towards Kevin Erdmann's brilliant take-down of the widely-accepted idea that there was a bubble in housing leading up to the 2008 meltdown. He does this through his own encyclopedic understanding of housing valuation and the housing market:

I know you are already aware of Kevin's work here, and in fact have blogged about it, but I wanted to point it out again, because it's such an awesome example of how people who don't understand asset pricing call "bubble" when the real issue is, well, they don't understand asset pricing.

I don't think anyone should be allowed to declare a bubble unless their personal balance sheet is net short by a significant amount.


Kenneth Duda
Menlo Park, CA

Dan W. writes:


Are you acquainted with the work of Vernon Smith? He conducted financial experiments to improve understanding of market mechanisms that promoted the development of "bubbles".

As some have pointed out already there was a tech bubble in 1999 - 2000. This is an empirical fact illustrated by a chart of the NASDAQ index from 1999 - 2003. The chart shows an accelerating price increase followed by a collapse with the price movement being far beyond "normal".

If the NASDAQ was not a bubble then what was it? To say it was a bubble is not to say that the behavior of market participants was irrational. Rather it is to observe that market participants each thought they could win the game of buying and selling better than others, no matter the price.

Yancey Ward writes:

If a bubble pops in a city, does it make a sound?

Tom P writes:

I think there would be a lot of legal restrictions on starting up the type of fund Scott proposes.

Based on my experience in the business, the vast majority of funds can't even short sell. And there are restrictions on leverage too. Most US investors don't want to hold foreign assets. And mutual funds are usually asset-class-specific to appeal to particular clienteles.

I think hedge funds do engage in the sort of trades you're suggesting, but there are legal restrictions that prevent most Americans from investing in these, and the case against hedge funds being able to beat the market is much weaker.

Ken P writes:

Maybe I'm wrong, but as I remember it the "bubble calls" in 2003 were people who said that implicit support for GSEs would distort markets and "result in" excessive home prices.

Also, if Fed money was flowing into the housing market due to QE purchases of mortgage backed securities, all things equal shouldn't we expect price distortion to result?

If someone wants to make bets against bubbles, they would have to simultanelously bet against future inflation and a wide range of potential government interventions as well as missing out on the significant short term gains of riding the bubble?

Does the phrase "X was overvalued at time T" have a meaning? If so, what is it?

@Dan W. You wrote "As some have pointed out already there was a tech bubble in 1999 - 2000. This is an empirical fact illustrated by a chart of the NASDAQ index from 1999 - 2003".

If the definition of a "bubble" is "prices went down", then, of course, bubbles exist. But, hey, the EMH, does predict that prices go up and then down.

1994-1998 was also highly unusual in the NASDAQ. Many people called it a bubble. But it wasn't.

Jason writes:


I think you really flub this one. Nothing you wrote above rebuts the idea that the "market can stay irrational" longer than you can stay solvent. The cost of maintaining a short position with stock Y with a price earnings ration of 400 can be very high and you can be wiped out before the stock crashes. It can go to 500 and beyond.
And Scott, You would have to be CRAZY to believe that P-E ratios of 500 or 600 are in any way justified fundamentally, even with the most phenomenal earnings growth rate imaginable.

And P-E ratios were in that range for tech stocks in the late 90's The only question was when the bubble pops. not whether bubbles exist!

Jose Romeu Robazzi writes:

Reproducing a comment I made over on "TheMoneyIllusion"

On Bubbles,
i have seen so many comments about “bubbles”, trying to figure it out using only prices. Readers of this blog and others have learned that the level of a price (interest rate) does not tell you anything about the monetary police stance. Moving to “bubbles”, valuations are hardly an indicator of “bubbles”. I have read many times that people cannot identify a bubble from prices, and the phenomenon of rapidly declining prices is associated with the idea of a “bubble” only after the fact, when it bursts. Now, the process which I think should be associated with a bubble is the amount of term mismatch in financing structures. It does not have anything to do with the debt level as well, but how it is financed. The investment banks went down in 2008 because they were highly levered, yes, but mostly because they were poorly financed, holding long term assets and financing it with one day repos. That caused the liquidation process after recongnized as a bubble bursting. I disagree with Sumner, I think bubbles exist, and yes, easy money is a condition that is necessary on the time frame before the burst, because easy money is one of the conditions that make term mismatch possible. But easy money alone does not “cause” a bubble. It just raises the probability one may occur.

Dan W. writes:

It seems to me that what trips up the ivory tower thinkers is the inability to make sense of the "animal spirits" of economic participants. Their solution is to assume such influence does not matter, if it exists at all. Keynes said otherwise, writing: "there is the instability due to the characteristic of human nature that a large proportion of our positive activities depend on spontaneous optimism rather than mathematical expectations"

Bubbles do not violate EMH and failed predictions of bubbles do not prove that bubbles cannot occur. It is an axiom that for every buyer of a stock there is a seller and each has a reason for engaging in the transaction. It is also an axiom that the future is unknown. Even in the midst of a bubble it is unknown how much higher a price may rise and it is unknown how low the price may be some distant time off. The only sure way to profit in the market is by making a trade that is immediately profitable: Thus the proliferation of HFT brokers. Any other trade that requires time to play out assumes the risk of loss.

The greater evidence of a bubble is the behavior of market participants. What are their justification for buying or selling and what is the financial structure of the transaction? Most important, how prepared are participants in the trade to deal with loss if the price moves against them? How long can traders “hold on” and will they have any reason to hold on? Those buying profitless dot-coms in 1999 had every reason to “hold on” as long as the price was moving in their favor. But when their position began losing value they had no reason to hold on as there was no reason to believe the company of the stock they held would ever have real, tangible value. The trade in these dotcom companies was pure speculation that became a mania that created an asset bubble that popped once the “animal spirits” changed.

The housing bubble was a bubble because the market activity was unsustainable. The marginal buyer did not have the cash flow to pay the mortgage without borrowing further against the house. But house prices were not going to increase indefinitely. Once prices stalled the economics that had sustained the housing market reversed and the financial structure that had propelled the housing market upward crumbled. The rational for buying a house at any price vanished and the “animal spirits” of market participants changed. Due to the entanglement of the housing economy and mortgage debt with the entire economy this reversal had great economic impact.

It is dangerous to assume away “animal spirits”. The wisdom, if there is to be any, is in minimizing entanglement of speculative and non-speculative financial capital such that losses are paid in full by those who make losing bets and never socialized to everyone else.

mobile writes:

To have an efficient market, investors need to be able to bet on their beliefs, including their belief that an asset price will decrease. This has always been pretty easy to do with stocks, but it's harder to do with real estate. For real estate bears, REITs and MBSs are an improvement over indirect methods like shorting construction stocks, though these are still not very popular vehicles. Maybe the housing markets were obviously mispriced in 2005, but the bears couldn't meaningfully participate in these markets until REITs and MBSs became more available in 2007-2008.

Scott Sumner writes:

Lorenzo and Rajat, Good points.

Rajat, I don't have an explanation of the NASDAQ in 2000, although each day that goes by it seems a bit less extreme.

Kenneth, Yes, I highly recommend that people look at Kevin Erdmann's posts on real estate.

Dan, I'm not sure how that comment relates to my post, which was on optimal mutual funds in a world of bubbles.

Tom, There are 100s of mutual funds with foreign stocks, indeed I own some of them. If mutual funds are unable to short stocks then the regulations obviously need to be changed. Why would such a regulation exist for mutual funds but not hedge funds? If you believe in bubbles (as most people do) then the law is incredibly anti-egalitarian.

Of course mutual funds could still invest in negative bubbles. Thus they could buy Japanese stocks when the yen is unreasonably low.

Another option is to invest only in non-bubble asset markets, that should outperform index funds.

Ken, QE money does not "flow into" markets. In 2003 publications like The Economist predicted that prices would drop in a half dozen housing markets. They were wrong in almost every case, but later claimed they had been correct. Indeed they bragged about it.

Daniel, Not to me, but it does to most people. I think to most people it means prices are higher than that particular person thinks is reasonable.

Luis, Good points.

Jason, I disagree on P/E ratios, and in any case you missed my point. I was talking about diversified mutual funds, not individual stocks. If bubbles exist, then any given year some bubbles will be bursting.

Jose, Bubbles do not typically occur during periods of easy money, just the opposite.

Hazel Meade writes:

The problem is that bubbles are a socio-psychological phenomenon, not really a financial one. Bubbles happen when consensus opinion feeds back on itself via market prices moving in response to trading activity that is itself based on consensus opinion. This causes something like confirmation bias, which reinforces consensus, which leads to prices moving in the same direction, this can continue until something dramatic changes the consensus view.
(Note that bubbles could be negative, an inverted bubble could be a crash or a bear market bubble).

The thing is that people who are going to call a bubble are by definitoin people who are outside the consensus view. In other words, the consensus opinion will ALWAYS fail to recognize the bubble, by definition. Because the bubble can't exist without shared psychological belief in it.

This makes it virtually impossible for society to do anything about bubbles. The government would always have to be acting on some minority belief rather than on the consensus view. And the consensus will never know when bubbles are happening.

Dan W. writes:


I was responding to an earlier comment where you wrote: "I don't believe that bubbles exist." (see comment dated April 25, 2015 10:33 PM)

Is this a belief you wish to defend? I get the impression that the position that matters to you is that it is difficult for a person to profit from a bubble. I agree. But this difficulty does not negate that bubbles occur. The post by Hazel Meade provides a good explanation of how and why.

Do you have any thoughts about Keynes and his "animal spirits". What about Vernon Smith and his work on experimental economics?

vikingvista writes:

The notion of a bubble as a psychological phenomenon is hard to defend. And every price rise with subsequent fall cannot be a bubble if prices are reflecting relative scarcities.

But the idea that relative prices are primarily changing due to something other than market scarcities, such as massive government fiat spending on securities, is not unreasonable. And given the disseminated nature of price information, you would not expect individuals to know whether or not existing prices are currently in such a bubble. All you would expect to know, is that such price manipulations will propagate in inconceivably complex ways.

What is hard to believe, is that such price manipulations don't exist or don't affect the production choices of people who experience them, or that they somehow wind up gracefully coordinated with less manipulated prices and real consumer demand.

Jose Romeu Robazzi writes:

@Scott, what I wrote does not contradict your view that bubbles burst during tight money periods.

Comments for this entry have been closed
Return to top