By: delta0618 on Giovedì 02 Dicembre 2004 10:58
Why I Think the S&P 500 Is Fairly Valued at 625
Many smart people, including most academicians, believe that the stock market is always fairly valued. Because the price is always appropriate, so the theory goes, there is no better or worse time to invest. In my opinion, that is bunk. The market's price wandered above and below its intrinsic value frequently during the last 123 years, and it is overpriced today. In fact, the S&P 500 looks dramatically overvalued based on two of the best methods for measuring market valuation.
Let me be clear: I can't "time" the market in the usual sense of that word. I don't know where it will go next month or next year. However, the success of stock-market investments should be measured not in months or years, but in decades. In the short term (less than 10 years), a cheap market can become cheaper, and an expensive market can become more expensive. Although the market might not reward you for paying attention to value in the short term, your long-term (30 years) returns will be determined by the price you pay. So how can you estimate the fair value of the S&P 500?
I think there are at least two valid methods for assigning value to the market: determining the "equity" Q ratio, and determining the cyclically adjusted P/E. Let's consider the Q ratio first. In Andrew Smithers' excellent book, Valuing Wall Street: Protecting Wealth in Turbulent Markets, he describes this ratio as the price of the stock market divided by its replacement cost. If the market is selling for less than its replacement cost, then entrepreneurs have an incentive to buy companies in the market because it would cost more to build them from scratch. If the market is selling for more than its replacement cost, then stock owners have an incentive to expand by issuing shares. The former action eventually drives stock prices back up to replacement costs, and the latter drives stock prices down toward replacement cost.
For the market as a whole (though not necessarily for individual companies) the Q ratio should equal 1. This is intuitively appealing and theoretically sound. There is one big problem--the replacement cost of the S&P 500 is difficult to measure, and there are a number of limitations in the data Smithers uses to create an estimate. Despite this, the Q ratio shows the property of mean reversion during the 20th century, and if investors had used it to guide investment decisions, they would have enjoyed higher returns and lower volatility than that produced by a buy-and-hold strategy.
The cyclically adjusted P/E (we'll call it the normalized P/E) overcomes the primary limitation of the Q ratio. Because earnings are volatile, a P/E ratio based on only one year of data is an unreliable indicator of value. By averaging earnings over a full economic cycle, we can estimate the true earning power of the market, and avoid being misled by an especially good or bad year. Robert Schiller, the author of Irrational Exuberance, figures that a decade ought to be long enough to encompass an economic cycle, so he adjusts earnings for inflation and deflation, and then calculates the 10-year average of real earnings. This creates the denominator for the normalized P/E.
In Valuing Wall Street, Smithers shows that the normalized P/E is equivalent to the Q ratio, but unlike the Q, the data for calculating a normalized P/E is readily available. It turns out that the market's fair value throughout the 20th century is about the same when determined by either method. Both measures revert to a mean, and because we know that the mean Q ratio is the theoretically correct fair value of the market, we can infer that the normalized P/E also defines fair value.
Because I am not certain that the imperfect data used to calculate Q will always exhibit the same consistent bias it has demonstrated in the past, I prefer to use the normalized P/E. On Robert Schiller's Web site, he provides enough data to calculate the fair value of the S&P 500 through March of 2004. The fair value then was about 606. Andrew Smithers provides an up-to-the-minute normalized P/E on his Web site, where it is labeled "proxy Q". Using this data, the fair value of the S&P 500 is about 625 today. You can also use Smithers' site to estimate fair value by using the Q ratio. As I write, that ratio is about 2, and the fair value of the S&P 500 is about 580. These figures are far below the recent market level of 1,180.
In case you are wondering, Morningstar's Market Valuation Graph shows the price to fair value ratio at 1.13 today--near its all-time high. However, the S&P 500 price is a market-capitalization-weighted average of 500 companies. Our graph represents the median price/fair value ratio for the nearly 1,500 companies that we follow. Therefore, the Morningstar market valuation graph is not comparable to the Q ratio.
In my opinion, it's a virtual certainty that the S&P 500 will revert to a "fair" Q ratio and normalized P/E. This doesn't mean that the market must drop. Its underlying value will rise over time. If the market price rises at a slower rate, then value will eventually catch up to price. As a practical matter, however, it's likely that the market will drop. The market's fundamental value changes slowly, but prices can fluctuate wildly. Throughout history, marked undervaluation or overvaluation has been corrected primarily by price changes. In a future article, I'll provide historical context for the market's current valuation. By knowing what has occurred repeatedly in the past, you'll be better prepared for what might happen in the future.
Wednesday December 1, 6:00 am ET
By Curt Morrison, MD, FACC