The first step in our process is to quantify—with certain critical modifications—the research that traditional analysts perform by hand. We use our own model to rank a universe of over 10,000 stocks according to expected investment return. The top 100-ranked stocks are large- and mid-cap companies with better-than-average profitability and long-term growth prospects selling at market or below-market multiples.

Our model does not consist of one simple screen or a combination of vendor-supplied screens; it is a truly proprietary approach to measuring expected investment returns. The model is the mathematical definition of expected return, rewritten in standard accounting variables. Specifically, expected return is stated as a function of three variables: (1) forecast growth, (2) forecast cash flow yield, and (3) forecast valuation change. For forecast growth, we use five-year normalized growth of earnings. Forecast cash flow yield is normalized return on equity in relation to the growth rate, which captures potential stock buybacks as well as dividends. Forecast valuation change represents our estimate of the expected future change in a company’s price/book ratio

The question often arises, "What numbers do you use?" We start with consensus Wall Street estimates, then modify the consensus to arrive at our own forecasts. This modification process results in consistently more accurate growth estimates, relative to actual results, than consensus estimates. In adjusting consensus estimates, we account for several quantitative phenomena documented by our research over the years:

> Earnings Overestimation Bias
Our studies have shown that analysts' earnings estimates tend to be significantly higher than actual earnings (“Falling in Love Again—Analysts' Estimates and Reality,” Financial Analysts Journal, September-October 1994). This overestimation bias has several potential sources: the very human tendency of analysts to "fall in love" with their stocks; the pressures placed upon analysts at investment firms that engage in investment banking activity; and the possibility that a company under analysis may sever the lines of communication in response to adverse coverage.

> Frequency and magnitude of
one-time charges
Recurring one-time charges can provide critical insights into future cash flows and the quality of company management (“One-Time Charges: Never Having to Say You’re Sorry?,” Financial Analysts Journal, September-October 1995).

> The impact of anomalous tax rates
Companies don’t pay unusually low taxes just because they have smart accountants.  They pay unusually low taxes because they really aren’t making money. We normalize tax rates when calculating earnings forecasts to arrive at normalized earnings capacity.

> Reversion to the mean
Companies with low price/book ratios trend upward over time, while companies with high ratios trend downward ("In Search of Excellence: The Investor’s Viewpoint," Financial Analysts Journal, May-June 1987 and "Excellence Revisited," Financial Analysts Journal, May-June 1994). We consider this phenomenon in estimating the third term of the model, the forecast change in valuation (price/book).

back to top