Baked in the Cake

How The Price You Pay Impacts Returns

Baked in the Cake

Probably the question heard most frequently is:“How high can stocks go?” To the unsophisticated observer there appears to be no maximum price.
New York Times, August 21, 1929

Investors must keep in mind that there's a difference between a good company and a good stock. After all, you can buy a good car but pay too much for it.
                                                             -Loren Fox,
Upside, July 6, 1999

When forecasting investment returns, individuals often make the mistake of simply extrapolating historic returns into the future. In addition, investors make the mistake of “recency.” Bull markets lead investors to expect higher future returns, and bear markets lead them to expected lower future returns. However, the price you pay for an asset has a great impact on expected returns. This bit of wisdom seemed to have eluded Henry Blodgett, Merrill Lynch Internet analyst, when he made the following statement in a report on Internet Capital Group: “Valuation is often not a helpful tool in determining when to sell hypergrowth stocks.”(1) He made this statement on January 10, 2000, shortly before the Internet bubble collapsed.

A two-step process determines equity prices. First, future earnings are forecasted. Then, the present value of those earnings is calculated by discounting them at the risk-free rate (the rate on riskless short-term instruments such as a three-month treasury bill) plus a risk premium (the size of which is commensurate with the amount of perceived risk). The lower the riskless rate or the risk premium, the higher the present value, and vice versa. Let’s explore how this works.

If the risk premium of an asset class falls (as investors perceive less risk), two things occur. First, investors in that asset class benefit from a one-time increase in the price of the asset as future earnings are now discounted at a lower rate. This is similar to the impact of falling interest rates on bond prices. The second impact is on future expected returns. Since risk premiums are a reflection of future expected returns, the falling risk premium reflects lower future returns. Of course, the reverse would be true if the risk premium of an asset class rose, as investors perceive greater risk. The first impact would be a drop in equity prices as future expected earnings are now discounted at a higher rate reflecting the now higher risk premium. The second impact would be that investors would now receive greater expected future returns reflecting the greater risk premium. This process is exactly the opposite of what investors perceive when they extrapolate the recent outperformance of an asset class into the future.

Let’s examine some of the historical data to see if we can make any useful observations about the size of risk premiums and future expected returns, or at least the changes in them. The average historical price to earnings (p/e) ratio for the market has been around 15. For the period 1926 through the second quarter of 1999, an investor buying stocks when the market traded at p/e ratios of between 14 and 16 earned a median return of 11.8% over the next ten years. (2) This is remarkably close to the long-term return of the market. The S&P 500 returned 11.0% per annum for the 74-year period 1926-2000.

Let’s now look at the returns investors received when they bought stocks when the perception of risk was low, such as during good economic times and a bull market; and when the perception of risk was high, such as during a recession or financial crisis and a bear market. Investors purchasing stocks when the p/e ratio was greater than 22 (when investors are highly optimistic and there is great enthusiasm for buying stocks) earned a median return of just 5% per annum over the next ten years. (3) High p/e ratios generally reflect a strong economy and a bull market. During such times, investors perceive relatively low levels of risk, which translates into high prices and low risk premiums. Those low risk premiums, however, also translate into low future expected returns - exactly the opposite of what most investors expect.

Let’s now look at the returns investors received when they purchased stock when the perception of risk was high. Investors who purchased stocks when p/e ratios were below 10 earned a median return of 16.9% per annum over the next ten years. (4) Low p/e ratios generally reflect a weak economy and a bear market. During such times, investors perceive relatively high levels of risk, which translates into low prices and high risk premiums. Those high risk premiums, however, also translate into high future expected returns. Investors buying stocks when the p/e ratios were below 10 (when perceived risk was high, and seemingly no one wanted to own stocks) outperformed investors that bought stocks when p/e ratios were above 22 (when perceived risk was low, and seemingly everyone was jumping on the equity bandwagon) by almost 12% per annum.

In that light, it is worth noting that from 1995-1999 we experienced a collapse in the risk premium for the large-cap stocks that dominate the S&P 500 and Nasdaq 100 indexes. At year-end 1994, the p/e ratio for the S&P 500 was just under 16, not much higher than its historical average. However, by the end of the first quarter of 2000, it had risen to just under 30, well above the 22 p/e ratio that has historically produced 5% returns over the succeeding 10 years. The Nasdaq 100 was trading at a p/e ratio of well over 100. Note also that the 1926-1994 return of the S&P 500 was just 10.2%. It took the bull market of the late 1990s (and the collapse of the risk premium) to raise the rate of return to 11%.

There have been very few episodes when p/e ratios have been as high as they are today. It seems “bubbles” only occur every generation or so - just long enough for those that experienced the pain to forget, and also long enough for a new generation of believers in the mantra of “this time its different” to become of investment age. However, the limited evidence we do have and, as you will see, the logic, suggests that the outcome is highly likely to be a period of either very low or even negative returns for the asset class of large growth stocks.

Unfortunately, investing is not a science. We don’t have clear crystal balls. This is best we can do is to put the odds in our favor. With that in mind, let’s examine four possible scenarios and their likely outcomes.

A) The risk premium (implied by p/e ratios) for large growth stocks reverts to its historical mean. This implies very poor returns for large growth stocks, as the p/e ratio would have to fall from its level of 28 on June 30, 2001, to its historic average of about 15. This implies either a very sharp short-term correction, or a long period of very low returns, until earnings can “catch up” with prices.

B) The risk premium remains stable. In this case returns are likely to be not
much higher than those on basically riskless TIPS, whose current real yield
is about 3.5%. We can estimate the returns for large growth stocks by
taking the earnings yield (E/P ratio), currently 3.6%, and adding in an
estimate for inflation. This can be done by taking the yield on long-term
bonds and subtracting the real return to TIPS. Today that spread is about
2%. Thus the expected nominal return to large growth stocks is only about
5.6%, only about 2% greater than the return to TIPS if inflation is zero.
If we assume a historically low level of inflation of say 2%, then the risk
premium for large growth stocks will be about 0. In addition, investment
grade corporate debt currently provides a higher yield than the expected
returns to large growth stocks in this scenario. This is an argument for
scenario A to unfold - if the risk premium is too small, it has a tendency to

C) The risk premium falls further. While a possibility, it also seems to be a highly unlikely scenario. While there really is no limit to how high the risk premium can rise, there is a limit to how far it can fall (excluding the possibility of irrational bubbles). The expected returns to equities should not fall below the expected return to riskless instruments. And, as we have seen, there is not much room to fall. However, a small fall in the risk premium would allow returns to be slightly higher than we are projecting. However, this still does not produce a very good outcome.

D) Corporate earnings grow faster. However, the problem is that there is no evidence that our economy we will grow any faster than it has historically. There have been many other periods of tremendous technological innovation and yet none resulted in much faster long-term growth. We have experienced the Industrial Revolution, and revolutions from the inventions of electricity, flight, television and radio, computers and semiconductors, etc., and none led to much faster long-term growth than the figure we are using to forecast returns. The Internet revolution is not likely to be any different. However, let’s see what happens if growth is greater than we are projecting. Since real GNP growth and real earnings growth are linked in a pretty consistent long-term one-for-one relationship, if earnings are growing faster so will GNP. Since there is a very close relationship between GNP growth and real interest rates, if GNP grows faster then real interest rates will also rise. That will drive the risk-free rate on competing instruments like TIPS and short-term fixed income rates higher. So the equity premium might not rise at all. So, while you would get higher equity returns, there would not be any extra return over risk-free alternatives.

The bottom line is that it is very hard to see a scenario of high returns for large growth stocks when prices being paid for future earnings are so high. Is it possible that returns will be greater than forecasted? Yes. But investing, not being a science, is about putting the odds in your favor. And, when it comes to equities, given the current high valuations of U.S. large growth stocks, diversification across asset classes, and not having all your eggs in one basket, is more important than ever.


(1) Smart Money, June 2001.
(2) Fortune, August 16, 1999.
(3) Ibid.
(4) Ibid.