About five years ago, as investors licked their wounds in the wake of the Enron and WorldCom scandals, analysts clamored for financial measures beyond earnings that might paint a truer picture of a company’s intrinsic value. Earnings, many argued, could be more easily inflated than other financial variables.
At the time, as had happened cyclically many times before, institutional investors rallied around cash flow from operations as perhaps the best sign of a company’s future cash flows and, by extension, its likely stock price.
Because cash flow, unlike earnings, excludes relatively arbitrary items like depreciation and amortization, the logic went, cash flow might be a closer proxy of a company’s value. Earnings in some high-profile scandals had proved fallible, since companies could fudge the numbers. Forecasts of other variables, including dividends, might also more perfectly signal a company’s value, some thought.
To see if, in fact, “Cash is king,” Professors Doron Nissim, Jacob Thomas of Yale and Jing Liu of the University of California, Los Angeles, studied financial data from I/B/E/S (Institutional Brokers Estimate System) for nearly 26,000 international companies from 1987 to 2004.
Studying variables from individual industries rather than all industries together and extending the analysis to other forecast variables by drawing on a larger swath of geographies and industries might make results more comparable and representative, they hypothesized.
“Theoretically,” explains Nissim, “forecast cash flows should be a better indicator of share price than earnings, since net asset value equals the present value of expected cash flows. And price is the present value of expected dividends.”
A 2002 study by the professors studied a sample of U.S. companies. That research showed that, on average, forecast earnings trumped all reported variables — including earnings, cash flow, sales, EBITDA and book value — as an indicator of a company’s share price, with sales and operating cash flows correlating least with a company’s share price.
The current study extended the researchers’ earlier work by looking outside the United States to a sample of listed firms in Australia, Canada, France, Germany, Hong Kong, Japan, South Africa, Taiwan and the United Kingdom. Using global data was important because analysts in some countries tend to calculate forecast variables beyond earnings more often than their U.S. peers.
International data would also allow for a more diverse group of industries and, by extension, more cross-industry comparisons. That is because nonearnings forecasts in the United States might be skewed, since they tend to be restricted to industries like mining, and oil and gas, with larger-than-average arbitrary expenses. Analysts following these industries often direct investors to variables beyond earnings like cash flow and dividends, which exclude these expenses.
Nissim and his coresearchers found that, on average, forecast earnings beat other variables — actual and projected — as a sign of a company’s stock price. The key takeaway for an underwriter pricing stock in a company about to go public or an investor calculating a company’s worth following that IPO? “Work with earnings,” says Nissim. “They may be a better indicator of a company’s value than cash flow.”
The researchers also found that forecasts nearly always beat their reported peers as an indication of value. Those results bore out across geographies and industries and when cash flow and dividends were also analyzed. In fact, projected earnings, the researchers say, were “remarkably accurate” measures of value for a “substantial majority of companies.”
And nearly always, forecast dividends outperformed actual dividends by only a small margin. That is perhaps because dividends tend to be “sticky,” or vary little over time. (The incentive to change the amount is minimal because most tax laws favor capital gains over dividends.) The exceptions were in Hong Kong and Australia, where dividends and capital gains receive the same tax treatment. For that reason, in those countries dividends tend to be a close proxy for company value.
The researchers also concluded that current and forecast dividend data appear to be less important in countries and industries with earnings forecasts that perform relatively well.
There are, as always, a few caveats. This analysis does not extend to companies with nonpositive earnings, cash from operations or dividends. Also, for early-stage companies in, say, the biotechnology or technology industries, other variables might be better indicators of value than forecast earnings.
Doron Nissim is professor of accounting at Columbia Business School.