P/E Expansion and Contraction
Last week, I showed a 50 year chart of the S&P500, focusing on the P/E over that time period. Today's chart covers the same issue, only we focus on the 1982-2000 Bull market.
Some people argue that P/E expansion wasn't all that significant; they say it was (if anything) a function of falling interest rates. In my mind, that only partially explains why multiples expand; It certainly cannot rationalize why P/Es went from 7 to nearly 50 over the course of ~20 years.
Why might the median P/E have run from 7 to 32 during the Bull market?
My explanation is Psychology: something shifted in investor sentiment that made them willing to pay more than $7 for a $1 of
earnings -- much more. That change is best explained by a sentiment shift related to perceived relative Value.
Click for larger graph
Most investors do not think P/E expansion as the lion's share of the market's 82-2000 gains; Instead, they credit a robust economy, technological advances, productivity gains, and (of course!) earnings improvement.
And all those elements did have an obvious impact -- by my math, they were responsible for about 25% of the performance.
But the biggest contribution these four elements had was not to the bottom line; rather, it was to investor psyches that gradually became willing to spend more per dollar of earnings than they had been. They allowed a rationalization of higher prices: Aren't stocks worth more if the economy is doing well? Doesn't technology make companies more efficient with their capital? If workers are more productive, went the thinking, than earnings will be all the more better.
Notice how squishy these thoughts are; they may be rational, but they are hardly the sort of easily quantifiable data points that makes for a dispassionate, calculating investor.
more on this later this week . . .
TrackBack URL for this entry:
Listed below are links to weblogs that reference P/E Expansion and Contraction:
In re, "a function of falling interest rates," it's worth doing a cocktail-napkin analysis just to check. Consider the old ddm, where P/E is payout*(1+g)/(r-g). 'g' estimates for the market will probably proxy for GDP growth, which is running around 3-4% in real terms. 'r' estimates are probably real treasury rates plus a risk premium; risk spreads are low right now, so r is probably not all that far from g. When your P/E is proportional to the inverse of r-g, and g is close to r, then P/Es will indeed be very, very sensitive to rates.
Like some out there, I tend to believe risk spreads and long US rates are both somewhat too low, which means that I agree with your P/E call. However, I do feel it's mostly a function of falling long rates, and of the associated drop in risk premiums.
Posted by: wcw | Jan 2, 2006 3:05:24 PM
The comments to this entry are closed.