IS THE STOCK MARKET OVERVALUED?

 
 
 " The problem of inaccurate data does not mean that the use of models in 
 financial decision making is undesirable. Without such models there would 
 be no means to value an asset. Hunches, intuition, or just plain guessing 
 would then be used to value and select assets. By using theoretical models,
 the portfolio manager is forced to identify real economic forces 
 (e.g., earnings and growth rates) and alternatives (e.g. the risk-free rate
 and the return earned by the market as a whole). Even if the analysis may 
 sometimes be inaccurate, it is still fundamentally sound and should prove 
 better than random guessing or intuitive feelings. "
 
 
 Herbert B. Mayo (1997)
 

Lester Thurow (1983) notes that economics is both a descriptive and normative science. Economic as a normative and mathematical science states how rational people ought to act, in a world where knowledge and markets are perfect. This is the economics of equations, where the coefficients are unspecified.

Economics as a descriptive, empirical science encounters many of the major difficulties that occur in the other social sciences. Andre Orlean (1999) mentions four major obstacles to developing economic laws - we have cited the examples: 

1) The question of mechanisms. When there is a general effect, it is difficult to know beforehand what sign will result when there are opposing causes. If the Fed increases interest rates, will that decrease the stock market? It also depends upon investors' assessments of the existing and future states of the economy.

2) The second problem he cites is the role of history, itself modifying institutional structures. On the issue of history, take an econometric data series that also includes the OPEC crisis. Now, simply delete those years. Since the statistical model is assumed to be stationary, the deletion merely reduces the statistical significance of the equation by reducing the number of cases. In fact, you have deleted some events that resulted in a restructuring of the entire world economy.

3) The third problem, Mr. Orlean states, is also fundamental. The essential difference between the social and natural sciences is that people learn, they modify their behaviors and beliefs with increasing knowledge. The announcement of a discovered economic law therefore changes the behavior of the economic system.

4) Individuals purchase goods because they are useful, for sustenance, communication, and so on. However, human actions depend upon social factors that economic theory only rarely apprehends, such as confidence, relationship, hierarchy, or moral code. Keynes, an astute investor, emphasized the importance of the general level of business confidence.

To ask whether or not the U.S. stock market is overvalued is to ask whether it is overvalued according to some quantitative determination of value. Having pointed out the obstacles to making such a determination, we do so using both normative and empirical models.

The Normative Model

The normative model of valuation states that the present value of an asset is the discounted sum of future benefits, discounted at the required return of the investment. Here are the levels of the S&P 500 calculated according to the varying assumption of dividend (earnings) growth:

Dividend Growth Rate vs. Calculated S&P 500

Growth Rate	Calculated S&P 500	% Overvaluation	Comments
7.4%	1473
7.3%	1359	2%
7.2%	1262	10%
6.2% (New Era)	731	89%	Productivity growth +1% over historical.
5.2% (Historical )	512	170%	Growth in EPS (1926-1999).

           (Actual S&P 500 8/25/99 = 1383, Dividends = $16.60)

In this analysis, we have discounted future dividends at a rate of 8.61%, the current long term yield of the AA telephone bond plus a 1% equity risk premium. We have chosen a rather low risk premium that favors stock investments. We cannot justify choosing a lower discount rate than the AA bond yield, because bonds are contracts and equities are not. Due to the effect of infinite compounding, the normative model is extremely sensitive to slight changes in the growth or discount rate assumptions.

Using historical data, even data adjusted for a "New Era," the U.S. stock market is currently very overvalued. 

An Empirical Model

Due to the problems of specifying both growth and discount rates, we prefer our empirical model that does not require these assumptions:

 
 Bond Yields(t)
 ------------ = .021 * cap util(t+1)- .071 * infl(t+1)
 Stock Earnings Yields(t)
 

The calculated value of the S&P 500 = 1038. By our model, the stock market is 34% overvalued.

By any reasonable measure, the U.S. stock market is overvalued. We now ask how this overvaluation is relevant. 

In 1989, the financial economist Robert Shiller published a book titled "Market Volatility." The thesis of his book is that the stock market is much more volatile than justified by economic theory. We reproduce his plot of the actual inflation adjusted P/E of the stock market (between the years 1901-1988) versus the P/E calculated by discounting future real dividends and the 1988 terminal stock price. The dotted line is the predicted value. The correlation Mr. Shiller found between the actual and predicted valuations was only .048.

Our empirical model has a much higher correlation with the fundamentals because of its specification and shorter time horizon, but financial models do not predict the effects of major unquantifiable events. After the fact, their error terms measure the impacts of these events.

                             Figure 4.2 by the kind permission of MIT Press, our emphases.

The stock market is much more volatile than a theoretical discount model predicts. In the language we use as value investors, the stock market can remain overvalued for very long periods of time. The stock market was overvalued between the years roughly: (1901-1913), the latter part of the 1920s, and (1955-1972). These were periods of increased economic predictability conducive to trade and investment.

At other times, it was undervalued or fairly valued: during the World War I, the 1930s, W.W. II, and the post-OPEC decade of the 1980s. (The P/Es that Mr. Shiller calculated are much higher than those normally utilized because he divides both theoretical and actual real (that is inflation adjusted) stock prices by the 30 year moving average of real past earnings, for econometric reasons.)

Within a given economic environment, the reasons for large and continuous decreases in valuation are related to specific major events, either financial panics unremedied by the monetary authorities or large economic readjustments, with increased inflation a symptom.

Reasons for Large Valuation Decreases (1901-1988)

Year	Reasons	Inflation Rate
1907	Financial panic.	n.a.
1929	Financial panic, attempts to restore a prewar gold standard (Temin, 1989).	n.a.
1937	The New Deal tightens monetary policy.	3.1%, 1.2% the previous year
1946	A large economic readjustment following W.W. II.	18.1%, 2.2% the previous year
1966	An inappropriate expansionary fiscal policy, causing an inflation that was also one of the causes of OPEC price increases.	3.3%, 1.9% the previous year

Keynes (1935) emphasized the importance of business confidence in addition to rational calculation; this idea of confidence goes beyond a simple mean-variance interpretation: 

  
 "The state of long term expectation, upon which our decisions are based, does 
 not solely depend...on the most probable forecast we can make. It also depends 
 on the confidence with which we make these forecast - how highly we rate
  the likelihood of our best forecast turning out quite wrong (our emphasis)...
 The state of confidence, as they term it, is a matter to which practical men 
 always pay the closest and most attention."
 

Will low inflation and high growth continue? Asked this question in a 7/99 PBS interview, former Treasury Secretary Robert Rubin said that no one knows. But most significantly, he suggested three conditions he thought were necessary for this growth to continue. We paraphrase:

1) The U.S. should not turn its back on the world and become isolationist. Economically, the government should maintain fiscal responsibility.

2) Businesses should attend to productivity.

3) The educational system should train its students well.

If these conditions persist, we think the U.S. stock market is likely to stay overvalued, with short-term fluctuations related to Fed actions. What does this mean for investment policy?

1) Due to overvaluation, the profit potential of the stock market is likely to be less than in the past few years. This means the truism that your asset allocation ought to be appropriate to your circumstances and risk tolerances is especially true.

2) Since financial markets and societies are complex, we don't think that market timing or excessive trading are effective portfolio strategies. Your investment strategy should offer a realistic possibility of adding value over the long term.

3) We try to get companies that can sustain their earnings growth, at the lowest possible prices, if and when they occur. The goal of Horizon Capital Research is to outperform the S&P 500 with stocks; we reiterate this with our earlier mentioned caveats. Risk control - a separate concern, and often confounded - is an asset allocation issue. It is very useful to distinguish between these two goals.

We conclude this article noting the relevance of financial models to investing. Because markets lack a marked tendency towards mean reversion, financial models are mainly a consistent method of doing things; informative rather than determinative. The overall valuation of the U.S. stock market changes according to the real issues of war and peace.

RETURN TO “THE STOCK MARKET, SCIENCE, AND HISTORY”
 

RETURN TO COMMENTARY AND REVIEW