Print this article | Return to Article | Return to CFO.com
Why the 80-year-old DuPont model still has fans.
Robin Goldwyn Blumenthal, CFO Magazine
January 1, 1998
Much has been said in favor of new financial metrics that marry such elements as growth rate and stock price to a company's inner financial mechanisms. Like old friends, however, some seasoned metrics continue to deliver value if only by virtue of familiarity and simplicity. The most prominent among these is the DuPont system of financial analysis. Developed in 1919 by a finance executive at E.I. du Pont de Nemours and Co., of Wilmington, Delaware, in the days when the chemical giant cooked up financial formulae as well as hydrocarbons, the DuPont system still helps many companies visualize the critical building blocks in return on assets and return on investments.
"The DuPont model is a way of visualizing the information so that everyone can see it," says accounting and MIS professor Stephen Jablonsky of Penn State University. A typical DuPont chart resembles a chart drawn to mark the progress of competitors in a tennis or basketball tournament. Entries that ultimately make up before-tax return on investment include cost of goods sold, selling expenses, administrative expenses, inventories, accounts receivable, and cash. At successive stages, they are added, subtracted, divided, or multiplied until return on equity is reached.
In his new book, The Manager's Guide to Financial Statement Analysis, to be published in February by John Wiley & Sons, Jablonsky stresses DuPont's enduring appeal. "Where financial people are integrated into the business," says Jablonsky, "they use traditional measures of evaluating performance, such as profit margins, return on assets, and return on equity, all contained in the DuPont model."
Outside the cloistered confines of finance departments, where financial performance has not taken root among rank-and-file workers, the DuPont model can be very effective. DuPont analysis "is a good tool for getting people started in understanding how they can have an impact on results," says Doug McCallen, budgets and forecasts manager at construction and mining equipment maker Caterpillar Inc., in Peoria, Illinois, where performance has turned around since the company launched a reorganization in 1991. That was the year it formally heightened focus on specific accountabilities for different parts of the business. "The DuPont model supports and reinforces those accountabilities," says McCallen.
Using return on assets as a primary performance measure, and setting targets for each part of the organization, Caterpillar can determine when a problem is related to operating efficiency versus asset utilization. The results have been phenomenal, McCallen says. ROA has reached double digits, accompanied by record profits in 13 of the past 15 quarters, as of last September 30.
Nucor Corp., a $4 billion-a-year producer of steel and steel products, has been using return on assets employed to measure performance in its facilities for years, with salutary results. "Number one, it's simple," says Sam Siegel, vice chairman and chief financial officer of the Charlotte, North Carolina-based company. "All of our people understand it." Each facility aims for a 25 percent or greater return on assets employed, which is calculated before the effects of federal income tax, interest expense, and corporate overhead.
Thoughts from time to time about trading DuPont in for something else have yet to convince Siegel that something else is better. Alternatives, in his view, won't improve the company's incentives program. Nucor's overall profit-sharing plan is based on a 10 percent share of companywide pretax profits, and Siegel notes proudly that Nucor has consistently paid that amount into the plan since 1966, a year after new management took over after Martin Marietta sold its stake in the company.
ROOM FOR ERROR
The performance model that a company chooses should depend on its culture, Siegel contends. Lacking that affinity, models with extra bells and whistles will not fill the gap. Not so, say proponents of economic value added (EVA), who argue that the traditional ratio-based DuPont model suffers from a serious, if not fatal, flaw. They charge that guesswork inherent in generally accepted accounting principles leaves room for managers to make shortsighted decisions that boost return on investment without real improvement in the business. "Ratio measures can be very misleading and create conflicts between management and shareholders," warns Dennis Uyemura, senior vice president in charge of financial institutions at Stern Stewart & Co., the consulting firm that developed the EVA framework in 1983. EVA has won popularity in corporate circles by correlating favorable results with decisions that produce returns exceeding the cost of capital--a formula for creating shareholder value. It also shifts such costs as research and development from the expense category to capital investment.
Gary John Previts, a professor of accountancy at Case Western Reserve University's Weatherhead School of Management, in Cleveland, does not agree that DuPont is significantly more vulnerable to manipulation than newer methods. "Some would say the income behind return on investment is an estimate and not a fact, because income requires judgment," says Previts. "Depreciation is a guess and accrual-based measures are somewhat guessy." But by the same token, he observes, "people who believe in estimating cash flows have to estimate them in the future, and who can estimate that with precision? It involves subjective guessing, as well."
Professor Jablonsky also cautions that all models that require numerous calculations and adjustments create opportunities for manipulation. At Caterpillar and Nucor, subjects of research projects by Jablonsky for the Financial Executives Research Foundation (the research arm of the Financial Executives Institute), methodology takes a backseat to clarity. "Performance information is widely shared," Jablonsky says, "and it's tied to compensation, so people take an interest in the numbers."
From an operating standpoint, managers may not need to know how well they are utilizing capital from day to day, according to Dale Flesher, Arthur Andersen professor of accounting at the University of Mississippi, Oxford. "Under the traditional measure, managers can't know whether they've created value for the firm, not without knowing the cost of capital," Flesher observes. "The question is, Do they care?" Without control over the company's cost of capital, he argues, a divisional manager may need to know only that the hurdle is 15 percent, for example, leaving the rest of the calculations to the corporate level.
This issue spawns sharp disagreement with proponents of cost-of-capital-based metrics, however. "One of the beauties of EVA is that the message is easily communicated to all levels of the business," says Kenneth Stephens, general auditor at Harnischfeger Industries Inc., a capital goods manufacturer based in Milwaukee. "With EVA, you are trying to make prudent business decisions to maximize your financial returns while minimizing the assets employed to achieve that. So you can talk to a salesperson about why you want to put consignment inventory in place, or ask a foreman whether he really needs a machine tool."
MIX 'N MATCH
The DuPont model falls short, say some critics, because it is not a great tool for predicting the future or for tracking costs. "It is designed to give you a picture of where you've put your resources," says Jack Kreischer, president of the Pennsylvania Institute of CPAs and managing director of Kreischer, Miller & Co., a public accounting firm in Horsham, Pennsylvania. A DuPont model can tell you, for example, the extent to which you've invested in machinery to build widgets, but it offers little insight into growth prospects or costs along the way, Kreischer says. By the same token, DuPont lacks the means to include increasingly prominent intangible assets in its return calculations.
Flexibility, a hallmark of the expandable DuPont model, can enable finance executives to combine return on investment with measures that do incorporate growth prospects. "DuPont can be hooked up to a variety of value-based measures," says Eric Olsen, vice president of The Boston Consulting Group Inc., which has developed cash flow metrics of its own, including cash flow return on investment, or CFROI.
Some securities analysts routinely use DuPont in their state-of-the-art environment. "The DuPont method has the advantage of being much easier to use than other forms of analysis that depend on information buried in footnotes," says analyst Jim Stoeffel of Salomon Smith Barney, in New York.
Nucor's decision to stick with DuPont reflects a pragmatic view of metrics. "I don't say it's the best way," Siegel concedes. "All I say is that it has worked for us."
Robin Goldwyn Blumenthal, a former reporter for the Wall Street Journal, writes frequently about business.
THE FIRST REENGINEER
------------------------------------------------------------------------ ------------ The DuPont model of financial analysis was dreamed up by F. Donaldson Brown, an electrical engineer who joined the giant chemical company's treasury department in 1914. A few years later, DuPont bought 23 percent of the stock of General Motors Corp. and gave Brown the task of cleaning up the carmaker's tangled finances--in effect, America's first large-scale reengineering effort. Much of the credit for GM's ascension afterward belongs to Brown's systems of planning and control, according to no less an expert than Alfred Sloan, GM's legendary chairman. Ensuing success launched the DuPont model to prominence in major U.S. corporations. It remained the dominant form of financial analysis until the 1970s. --R.G.B.