Savvy executives know that the decision to invest in a project often hangs on reasonable expectations of its return on invested capital. But what constitutes “reasonable”? Companies that rely on the wrong benchmark can overlook good investments or pursue bad ones. We find that empirical analyses of ROICs—particularly those illustrating industry-specific patterns over time—can help executives ground their expectations in the collective long-term experience of other companies.
We analyzed the ROIC histories of about 7,000 publicly listed nonfinancial US companies from 1963 to 2004. These companies had revenues of more than $200 million in 2003 dollars, adjusted for inflation. Our sample included active companies as well as companies that were acquired or dissolved, and we looked at patterns that both included and excluded goodwill. The revenues of the companies we studied account for 99 percent of those of all nonfinancial US publicly traded companies in 2004, or some 82 percent if financial ones are included. Our work had several key findings.
First, the average US company has returned its cost of capital over time. From 1963 to 2004, the US market’s median ROIC, excluding goodwill, averaged nearly 10 percent. That level of performance was relatively constant and in line with the long-term cost of capital (Exhibit 1). The stable median ROIC may reflect a balance between investment and consumption. Companies that drive innovations in technology or business systems may earn above-average returns initially, but competition eventually compels most businesses to pass the savings along to consumers.
Second, historical ROICs can vary widely by industry. In the United States, the pharmaceutical and consumer packaged-goods industries, among others, have sustainable barriers, such as patents and brands, that reduce competitive pressure and contribute to consistently high ROICs. Conversely, capital-intensive sectors (such as basic materials) and highly competitive sectors (including retailing) tend to generate low ROICs.
These differences in the way industries perform haven’t changed substantially over time. In Exhibit 2, the ROIC ranking, based on the ranking for 1963 to 2004 as a whole, largely mirrors the average for the period from 1995 to 2004. In general, the persistence of differences in ROIC across sectors suggests that individual companies should be benchmarked against comparable ones operating in similar or adjacent industries.
Finally, we found that the median or mean returns of general, broadly defined industry groups can be downright misleading. Executives who look at the mean or median ROIC of an industry without understanding the distribution of ROIC performance within it may not have sufficient information to assess a company or to project its ROIC accurately.
Indeed, intra-industry differences are sometimes far more dramatic than those among sectors (Exhibit 3). Take the software and services industry. Its median ROIC from 1963 to 2004 was 18 percent, but the spread between the top and bottom quartile of companies averaged 31 percent. In fact, the industry’s performance was so uneven as to render this metric meaningless. These wide variations suggest that the industry comprises many distinct subgroups that have very different structures and are subject to very different competitive forces. To form reasonable expectations, it is often necessary to dig down to more refined subindustry groupings. By contrast, the utility industry’s median ROIC is only 7 percent, but the spread from the best to the worst companies is a slim 2 percent. Any executive encountering projected returns outside those of this relevant benchmark industry range would do well to look on those forecasts with a gimlet eye.