Nothing about the investment climate is ever certain. Anyone involved in investing for a reasonable period has been humbled numerous times by the unpredictability of the market. So carefully assessing the odds of various investment choices is essential for the success of a portfolio. I have often heard claims that a stock is attractive because it offered 100% upside with only 10% downside. Such claims offer two possibilities, yet fail to establish odds for each outcome. Taken to the extreme, buying a lottery ticket for $1 offers the buyer a chance for multiples of upside with only 100% downside. Unfortunately, the odds of a favourable outcome are exceedingly small while the odds of a total loss are nearly equal to one. The expected return for those who repeatedly buy a lottery ticket is, in fact, negative.
For the purpose of accumulating wealth through the long term, the favourable odds of investing in stocks is well established. Professor Jeremy Siegel in his book, Stocks for the Long Run, offered these figures. Since 1802, $1 invested in the stock market would have become $7.47 million by 1997, while the same dollar invested in bonds only becomes $10,744. Furthermore, the same dollar sitting under a mattress would only have the purchasing power of 7 cents by 1997. Inflation gives one the incentive to pull the money out from under the mattress and stocks give one the reason to put the money to work.
It has been very rewarding merely to keep up with stocks’ average return. However, two factors, investing in low valuation stocks and stocks showing earnings surprises, have consistently beaten the stock market averages. Low valuation stocks are characterized by a low price-to-earnings ratio or a low price-to-book ratio. Book value is essentially the productive assets of a company while earnings are its productivity. So buying assets and earnings is the ultimate purpose of buying stocks. It is no surprise that paying less for assets and earnings offers a greater reward. Benjamin Graham and David Dodd first advocated value investing in their 1934 book, Security Analysis. The odds for a favourable return by investing in value stocks was later empirically demonstrated by numerous authors, culminating in Eugene Fama and Kenneth French’s 1992 publication in The Journal of Finance.
In the late 1960s, financial researchers first noticed that stocks of firms reporting good earnings showed good performance well after the announcement. However, it was not until the late 1970s that researchers began to notice the effect of earnings surprises. Good earnings were then quantitatively measured as earnings above a certain expectation. Today, after many more studies, the correlation between earnings surprises and subsequent stock performance is well established. Lawrence Brown, in the March/April 1997 issue of Financial Analysts Journal, comprehensively reviewed many of these studies.
In quantitative modelling, earnings surprises are often measured as the difference between the reported EPS and consensus EPS supplied by sell-side brokerage analysts. In recent years, a phenomenon called ‘the whisper number’ has gained momentum. These are revised earnings estimates in which brokerage analysts incorporate the most current business and economic conditions. These numbers are widely disseminated to large institutional investors. As a result, a company’s stock price often reacts to whether the company’s reported EPS is above or below this ‘whisper’ earnings estimate, rather than the more stale published EPS estimate. A natural question to ask is whether the effect of earnings surprises has been diminished due to the development of ‘whisper’ numbers. One must understand that the difference between reported and consensus EPS is merely a proxy for earnings surprises. It is used because it is easy to model. Earnings surprises ultimately represent the difference between reported earnings and investor expected earnings. In many cases, the ‘whisper’ number is now a better approximation of investor expectation. The challenge for financial model builders is to find better proxies for earnings surprises.
By combining the strategies of investing in value stocks and companies showing positive earnings surprises, we have enhanced our odds of outperforming the stock market. However, to translate superior odds to superior performance, we also must limit the variability of outcomes. Consider an asset with the equal probability of appreciating 50% in value and losing 20% of its value. While the expected return for this asset may be 15%, half the time one will still suffer losses. However, if we have an uncorrelated asset with an identical return profile, we can cut our loss probability to only 25% by investing half the funds in each asset. In portfolio management, cutting the odds of unfavourable outcomes is precisely the task of risk management.
Well-diversified portfolios with strategies of identifying and employing superior odds are common to many quantitatively managed products. However, most such products have not been very successful despite the favourable odds. Execution shortfall, fees and transaction costs are all part of the reason. I believe the lack of fundamental analysis to support the quantitative model is another crucial reason. The stock market offers many seemingly attractive opportunities that disappear upon closer examination. Quantitative models are indeed good at finding these.
For example, in the autumn of 2000, Californian utilities were reporting excellent earnings while their stocks traded at rock bottom valuations. But a closer look at their balance sheets and cash flow reveals that the so-called earnings only existed in an ‘accounting sense’. The loss accumulated by selling power was recorded as an asset on the balance sheet resulting in a time bomb waiting to explode. Other examples of phantom opportunities involved many high-flying stocks like JDS Uniphase and Nortel. In fact, both of these stocks are included in the new S&P Barra Value Index. On the surface, both stocks trade at very low price to book value. However, both companies have made numerous high priced acquisitions in the past and the majority of their book value is comprised of good will. These and other so-called opportunities can easily be eliminated by doing a little fundamental analysis. In so doing, a fund’s odds of beating the market are further enhanced.
Ten years ago, we set out to design a product that would consistently beat a broad- based benchmark. We formulated a strategy that ranked stocks’ attractiveness based on a combination of valuation and earnings surprises. We diversified our portfolio with over 200 stocks across all sectors and industries. We also insisted on conducting fundamental analysis on every purchase candidate. Today, we are happy to report that our quantitative strategy has outperformed the S&P 500 Index in seven out of the past 10 years.
Dong Hao Zhang and Steve Colton are managing directors and portfolio managers responsible for the Quantitative Equity product at Phoenix Investment Partners