Saturday, May 16, 2015

"On the Importance of Asset Class Bubbles for Value Investors and Why They Occur"

This was from a Jeremy Grantham speech at the Annual Benjamin Graham and David Dodd Breakfast (Columbia University, October 7, 2009). The entire thing is excellent, but his point about profit margins and P/E multiples is very important:

Every time the market crosses fair value, it’s efficient. For a few seconds every five or six or seven years, it’s efficient. The rest of the time, it is spiking up or spiking down, and is inefficient.

Now, the market should equal replacement cost, which means the correlation between profit margins and P/Es should be −1. Or, putting it in simpler terms, if you had a huge profit margin for the whole economy, capitalism being what it is, you would want to multiply it by a low P/E because you know high returns will suck in competition, more capital, and bid down the returns (conversely at the low end). But what actually happens? Instead of having a correlation of −1, our research shows it has a correlation of +.32. The market can’t even get the sign right! High profit margins receive high P/Es and vice versa, and the correlation is much greater than +.32 at the peaks and the troughs. Right at the peak in 1929, we had record profit margins and record P/Es. In 1965, there were new record profit margins and record P/Es (21 times). Now, think about 2000. We had a new high in stated profit margins and decided to multiply it by 35 times earnings, a level so much higher than anything that had preceded it. In complete contrast, in 1982 we had half-normal profits times half-normal P/Es (8 times). I mean, give me a break. We were getting nearly one-third of replacement cost at the low, and almost three times replacement cost at the high in 2000.

This double counting is, for me, the great driver of market volatility and, basically, it makes no sense.
In other words, coming to the same conclusion as Hussman but from a different angle.

No comments: