Monday, 13 August 2012

Cyclical Variations in Equity Prices

Below is an excerpt from an academic paper I wrote back in 2001 on cyclical variations in equity prices, the equity risk premium and mean reversion of equity returns. The paper has a focus on shares of cyclical companies, but is also applicable to the overall stock market, especially these days as we could very well experience significant fluctuations in the equity risk premium in the immidiate future in the U.S., Europe and other markets as well. And the risk is on the upside, an increase in the overall equity risk premium that is, caused by financial issues especially in the eurozone.
A feature of most economic time series is inertia, or sluggishness, and it is a well-known fact that time series such as GNP, price indexes and industrial production exhibit (business) cycles (Gujarati (1995)). This section reviews our philosophy that equity prices fluctuate more than the underlying value and the intuition and theoretical foundation underlying such fluctuations. We also believe fluctuations in equity prices are more prevalent for cyclical industries than for industries with more stable cash flows and economic prospects. The theory underlying this statement is as follows. Cyclical industries are highly exposed to the general business cycle, they often have long histories and face slow change and stable competition, and operate in largely demand driven industries. Such “boring” and “old fashioned” industries experience high fluctuations in profitability and gain double-digit profit growth in market booms and experience significant losses during recessions. With reference to the empirical research reviewed in the section about mean reversion (e.g. overreaction, naive extrapolation of past growth rates, contrarian investment strategies preferring value overgrowth and out-of-favour stocks over glamour stocks, and time varying equity risk premiums) the hypothesis of this research is that cyclical industries are thought to be the perfect prey of overreaction and extrapolation of recent past growth rates by investors one time or another on its ride through the business cycle. More specifically, the market will expect further growth when the company is in fact approaching the peak in profits (overly optimistic) and will expect even larger losses when in fact the company is close to the trough (overly pessimistic). This will cause overreaction at both extremes and lead to over valuation of shares at peaks and under valuation at troughs thereby inducing mean reversion. Negative serial correlation is expected to be more prevalent the more the market overreacts at both troughs and peaks. The theoretical underpinnings of this expectation is in agreement with Barsky and De Long (1990) who believe a case can be made that the market and its analysts and economists overreact, and that things are never as good as they seem in a boom or as bad as they seem in a depression.
Part of the expected fluctuations may be explained in terms of the CAPM beta, since cyclical companies by definition should have higher betas than firms with more stable cash flows (Brealey et al (2000)). However, Ball (1995) holds that finance theory says nothing about how betas vary among companies and over time. If one assumes that a company has a given level of overall risk, i.e. a fixed asset beta, the beta of the company’s shares will rise along with increases in its degree of financial leverage. Ball (1995) loosely interprets this as “when a company’s stock price drops sharply, investors’ perception of the stock’s risk and hence its required return actually increases. Conversely, when the stock price rises significantly, its perceived risk and required return fall”. This study argues that these time-varying fluctuations is not always justified, that they have irrational features and that a long-term efficient market should exhibit smoother trends in share prices and the equity risk premium. Indeed, evidence presented earlier shows that periods of high implied risk premiums represent lower risk, even though financial theory holds the opposite. Logically, this makes sense. Since the growth in dividends for the broader market is rather stable over time, periods of historically high equity prices suggest the downside risk is higher during such times. During historically low levels, the opposite is true. The overall implication for a true long-term efficient market is that the risk premium should be higher during bull markets and lower during bear markets compared to what they normally have been. This type of stock market would then exhibit along-term smooth trend with quick mean reversion to intrinsic values (See Lee, Myers and Swaminathan (1999)). Typically, the market will turn well in advance of actual changes in profitability and valuation often tends to lead stock prices (Kavli (2001)). For example, steel shares normally perform six to nine months before sustainable steel price increases are in sight (Lindh and Haugerud (2001)). The real question here is not when, but whether the market overshoots. In the long run, valuations are expected to revert to some long run average representing the intrinsic value, or more illustrative in this context, the central value. Share prices above this central value is then thought to represent temporary over valuation, and prices below are thought to indicate temporary under valuation. Such reasoning is in line with Shiller (1981) who concludes that stock price volatility during the period 1871 to 1979 for the Standard & Poor index (monthly returns) was five to thirteen times too high to be attributed to new information about future real dividends. He explains that it has often been discussed that stock markets are more volatile than what the underlying change in fundamentals suggest. This implies that ex post, the movements in share prices could not “realistically be attributed to any objective new information since movements in the price indexes seem to be too big relative to actual subsequent events. Similar findings regarding volatility are reported by Barsky et al (1993) who stress that dividends are smoother than stock prices and that share prices appear to overreact to long swings in dividends. Specifically, they report that a long-run 1 percent increase in the level of dividends is associated with an approximately 1.5 percent increase in equity values. This suggests that share prices indeed depart from the central value at times. This is further supported by Summers (1986) who argues that both theoretical and empirical considerations suggest the likelihood that market valuations differ frequently and substantially from fundamental values. He holds that a classic evidence of divergences between the market and fundamental valuations is the discounts on closed end funds, a case where the underlying assets are easily valued. Barsky et al (1990) make a case that these “excessive” movements could be justified since even small changes between the cost of equity and growth, hereafter referred to as the spread, can have sizeable effects on valuations. In the same spirit, Mehra and Sah (2001) find that small fluctuations in investors’ discount factors (caused by small fluctuations in investors’ subjective preferences and bias) induces large fluctuations in equity prices.
The findings from the above discussion and review of empirical findings culminate to an important observation: If dividends and interest rates are relatively stable through time (in the long-term), then it follows that dividend growth rates are relatively stable. If temporary shocks to dividends have little effect on long term value, this strongly indicates that the equity risk premium explains the majority of movements in stock prices. Security prices are therefore rational and markets efficient only to the extent that time-varying equity risk premiums are rational. Such time-varying risk premiums might be viewed as rational in an equilibrium model incorporating time-varying premiums. However, if high implied equity premiums consistently leads to high returns, in which case the risk in fact is lower than that implied by the premium, the markets’ long-term efficiency should be seriously questioned. If the latter holds true, the premium can be viewed as a commodity which should mean revert and become more stable through time in a rational market with a long-term investment horizon.

No comments:

Post a Comment