The Yield Yen : A Case History of the Importance of Other Systematic Risk Influences
Is the quant quest for better risk measures an assault on windmills – a useless exercise that succeeds only in enabling academics to continue to play with their computers? No! It has important implications for protecting investors. Take, for example, the yield yen. The yield yen, which attracted a considerable following in the investment community by the 1980s, was the idea that institutional investors should place their funds in a yield-tilted index fund.
The reasoning behind this proposal seemed appealingly plausible. Dividends are generally taxed more highly than capital gains. This was especially true before the Tax Reform Act of 1986. since the market equilibrium is presumably achieved on the basis of after-tax returns, the equilibrium pre-tax returns ought to be higher for stocks that pay high dividends than for securities that produce lower dividends and correspondingly higher capital gains, which, even if fully taxed at regular rates when realized, provide some tax benefits since unrealized capital gains do not bear any tax at all. Hence, the tax-exempt investor should specialize in buying high-dividend-paying stocks. In order to avoid the assumption of any greater risk than is involved in buying the market index, however, this tax-exempt investor is advised to purchase a yield-tilted index fund; that is a very broadly diversified portfolio of high-dividend-paying stocks that mirrors the market index in the sense that it is constructed to have a beta coefficient exactly equal to 1.0.
Even on a priori grounds one might question the logic of the yield-tilted index fund. The validity of the proposal rests on the premise that the major market participants prefer to receive income through capital gains rather than through dividends. But many of the largest investors in the market (such as corporations) actually pay a higher tax on capital gains than on dividend income (For corporate investors, 70 percent of dividend income is currently excluded from taxable income, while capital gains are taxed at normal gains rates). It is far from clear that the most important investors in the stock market prefer to receive income in the form of capital gains. Therefore, the market may not price high-dividend-paying stocks so that they offer especially attractive returns to tax-exempt institutions. But apart from these a priori arguments, the statistical results just reviewed can be interpreted as providing another argument against the yield-tilted index fund.
If the traditional beta calculation does not provide a full description of systematic risk, a yield-tilted index fund may well fail to mirror the market index. Specifically, during periods when inflation and interest rates rise, high-dividend stocks may be particularly vulnerable. Public-utility common stocks are a good example. Although they are known as low-beta stocks, they are likely to have high systematic risk with respect to interest rates and inflation. This is so not only because they are good substitutes for fixed-income securities but also because public utilities are vulnerable to a profits squeeze during periods of rising inflation, as a result of regulatory lags and increased borrowing costs. Hence, the yield-tilted index fund with a beta of 1.0 may not mirror the market index when inflation accelerates.
The actual experience of yield-tilted index funds during the period of their popularity in the early 1980s was far from reassuring. The performance of these funds was significantly worse than that of the market. At other times, high yield stocks have significantly outperformed the market. Of course, we should not reject a model simply because of its failure over any specific short-term period. Nevertheless, we believe that an understanding of the wider aspects of systematic risk, as analyzed here, can potentially help to prevent serious investment errors.
A Summing Up
The stock market appears to be an efficient mechanism that adjusts quite quickly to new information. Neither technical analysis, which analzyes the past price movements of stocks, nor fundamental analysis, which analyzes more basic information about the prospects for individual companies and the economy, seems to yield consistent benefits. It appears that the only way to obtain higher long-run investment returns is to accept greater risks – and those risks can be horrendous, as any investor who has lived through the great bear markets of the late 1960s and 1970s and who suffered through October 1987 can tell you.
Unfortunatley, a perfect risk measure does not exist. Beta, the risk measure from the capital-asset pricing model, looks nice on the surface. It is a simple, easy-to-understand measure of market sensitivity, and differences in long-run rates of return form portfolios are clearly related to that single risk factor.
Unfortunately, beta also has its warts. The actual relationship between beta and rate of return does not correspond to the relationship predicted in theory. Moreover, the relationship is undependable in the short run and has even failed to work in periods as long as a decade, such as the 1980s. Finally, beta is not stable from period to period, and it is sensitive to the particular market proxy against which it is measured.
No single measure is likely to capture adequately the variety pf systematic riisk influences on individual stocks and portfolios. The actual relationship between beta and rate of return does not correspond to the relationship predicted in theory. Moreover, the relationship is undependable in the short run and has even failed to work in periods as long as a decade, such as the 1980s. Finally, beta is not stable from period to period, and it is sensitive to the particular maket proxy against which it is measured.
I have argued here that no single measure is likely to capture adequately the variety of systematic risk influences on individual stocks and portfolios. Returns are sensitive to general market swings, to changes in interest and inflation rates, to changes in national income, and, undoubtedly, to other economic factors such as exchange rates. And if the best single reisk estimate were to be chosen, the traditional beta measure would not be the only possibility. The mystical perfect risk measure is still beyond our grasp.
To the great relief of assistant professors who must publish or perish, there is still much debate within the academic community on risk measurement, and much more empirical testing needs to be done. Undoubtedly, there will yet be many improvements in the techniques of risk analysis, and the quantitative analysis of risk measurement is far from dead. Our guess is that future risk measures will be even more sophisticated – not less so. Nevertheless, we must be careful not to accept beta or any other measure as an easy way to assess risk and to predict future returns with any certainty. You should know about the best of the modern techniques of the new investment technokogy – they can be useful aids. But there is never going to be a handsome genie who will appear and solve all our investment problems. And even if he did, we would probably foul it up – as did the little lady in the following favorite story of Robert Kirby of Capital Guardian Trust:
She was sitting in her rocking chair on the porch of the retirement home when a little genie appeared and said, ”I’ve decided to grand you three wishes.”
The little old lady answereed, “Buzz off, you little twerp, I’ve seen all the wise guys I need to in my life.”
The genie answered, “Look, I’m not kidding. This is for real. Just try me.”
She ahrugged and said, “Okay, turn my rocking chair into solid gold.”
When, in a puff of smoke, he did it, her interest picked up noticeably. She said, “Turn me into a beautiful young maiden.”
Again, in a puff of smoke, he did it. Finally, she said, “Okay, for my third wish turn my cat into a handsome young prince.”
In an instant, there stood the young prince, who then turned to her and asked, “Now aren’t you sorry you had me fixed?”