Tuesday, July 20, 2010

On Spotting Bubbles: Part 1 of Many

One of the most fascinating phenomena in financial markets is the "bubble". Defining a bubble is somewhat of a tricky art form. I think most people would agree with this quote by Justice Potter Stewart on the issue of obscenity: "I know it when I see it". We all think we know a bubble when we see it, but do we really? I have found myself looking for definitive methods for spotting bubbles before they are clearly out of control and I wanted to share some thoughts on the subject.

Broadly speaking, I would essentially define a financial asset bubble as a case where the price of an asset has far outstripped any reasonable relationship to what it should be priced off of. With stocks, this is the earnings power of a company. With housing, it is the price to rent ratio or a price to income ratio with some adjustments. Asset prices, no matter how much people may want to believe otherwise, must necessarily be constrained by their relationships to their fundamentals over long periods of time. If this was not a constraint, we could all become infinitely wealthy by speculating in financial assets. What these fundamental relationships should be (i.e. a PE of 15x vs 17x or a price to rent ratio of 1.1x 1.0x) are the province of markets to flesh out over any given time, but in the long run, certain reasonable relationships do hold.

"Ah", you might say, "You said in the 'long run'. What about the short run where I actually live?" Indeed, and this is the problem. Over short (3-5 year time spans) any particular asset class can become wildly overpriced. How wildly? Well, let's look at the NASDAQ in the span of 1996-2000:

As you can see, up until 1999, the NASDAQ held more or less in line with the S&P 500 before wildly departing and going absolutely bonkers (technical term). What's more is that the S&P 500 in this time frame was also overvalued quite considerably, though I will discuss that finer point in detail in later post. Earnings growth during this time span was good, especially for the technology-heavy NASDAQ, but not anywhere near that good. It's much like how ordering pasta at a high-end Italian restaurant may be worth $15.00 a plate, but not $30.00. At $30.00, you are just being irrational.

This takes me to a brief digression on economists. Many economists, particularly those with significant classical leanings (no, they don't sit around and listen to the finest works of Handel, though they might) feel that consumers and investors are always rational and they justify this point with a bit of twisted logic. This is that because an investor bought an asset at a certain price and expects to receive more for it in the future, they are being rational. Only if they bought an asset with the expectation that they would lose money in the future would they be irrational. If this definition of rationality strikes you as absurd, it should. If I were insane and kept drinking six cans of Dr. Pepper everyday under the expectation that would be good for my health, would that be rational? Well, it might be if I were poorly informed, but I would already have to be irrational to believe such a thing.

This brings me to how to, qualitatively, spot a bubble. You do not do so by saying "Oh, the stock market is up 50% over the past twelve months so that must be a bubble" because the market has done that plenty of times and not looked back ever again. Similarly, in individual sectors, a rapid advance does not necessarily mean that people are not in league with their senses. Steel production in the United States in the late 19th century increased many thousands of percent and did not constitute a bubble. Similarly, PC sales from 1985 to 1995 increased by such a large percent that most people wouldn't believe it if you printed it. As such, a simple quantitative rule such as "x% increase necessarily equates a bubble" is not useful.

The basic qualitative framework for assessing the likelihood of a bubble is trying to determine the amount of reliable information (broadly construed) available in a market and then assess to what extent that information is being employed rationally. There is a third variable that is unfortunately quantitative, but is easily enough assessed which is whether the asset or whole asset class in question present offer greater than typical rates of return. Bubbles do not form in assets or asset classes where there is not a truly better than typical fundamental story going on. For example, the U.S. housing bubble formed during a time of low inventories and high affordability due to low interest rates. The NASDAQ bubble occurred in the sector of the economy seeing the fastest growth.

So, to put succinctly:
1. Is the asset or asset class in question fundamentally more attractive than other alternatives?
2. Is there either little information available or is the information corrupted in some way?
3. Are market actors incorporating available information or are they doing so in a rational way?

As I hope to show in later posts, this basic framework can be used to assess not only the presence of bubbles, but also to determine the differential effects of bubbles within an asset class. Incidentally, these criteria can be used to assess what I call "fear bubbles" such as what existed in March 2009.Any source

No comments:

Post a Comment