Thursday, October 9, 2008

some historical perspective

Negative correlation could reflect "mean reversion", i.e., as interest rates fall relative to stocks, they become less attractive as an alternative so investors sell bonds and buy stocks, driving up the stock price and raising the P/E ratio.

But remember this is an equilibrium relationship between two variables--causality cannot be inferred. You could just as easily tell a story of the relationship with the causality reversed, e.g., technology increases future earnings so P/E ratios increase and interest rates fall as investors sell bonds and buy stocks.


  1. Wouldn't buying stocks raise the PE ratio?

  2. yes, my mistake. thanks. corrected in blog post.

  3. Don't you need to use real interest rates? P/E is real, so
    the interest rate must be real. That would certainly lower the late 70's/early 80's spike.

  4. I will try to get the series online--if you have it, I will post it.

    If earnings grow at a constant rate, then the Price=Earnings/(r-g) where r is the cost of capital and g is the growth rate in earnings.

    Rearranging this equation gives us
    E/P=r-g. If we think that r is close to the opportunity cost of capital and that g doesn't change very much over time, then the historical relationship between E/P and r should tell us something about stock value.