The Performance Puzzle
All firms aim for strong performance, but few worry about what comprises such performance. Perhaps they should. The current run [...]
April 1 1992 by Marshall W. Meyer
All firms aim for strong performance, but few worry about what comprises such performance. Perhaps they should. The current run up of the stock market, at a time when corporate profits are flat to declining, raises the question of whether corporations doing a good job for their shareholders, are, indeed, doing a good job. Clearly, some CEOs believe that shareholder value is all-or almost all-that counts. Consider this statement by Lloyds Bank’s CEO Brian Pitman, which appeared in the company’s 1989 annual report. “In 1989, Lloyds Bank reported a big loss. Yet it was another good year for our shareholders.”
Mr. Pitman’s encomium for share prices is not universally shared. Others are skeptical of the ability of securities markets to assess the performance of a firm accurately, at least in the short run. As John F. Childs, a Kidder Peabody vice president, wrote in Public Utilities Fortnightly.
“… a company can be performing very poorly and continue to perform very poorly but if there is an improvement, even though still unsatisfactory, the market price may increase and produce a high return to investors… Stock market performance… is not a proper measure of company performance… “
Whether Mr. Pitman or Mr. Childs is right is immaterial. The point is that they disagree, as many reasonable people will, as to what constitutes satisfactory performance.
It is widely understood that share prices and accounting measures, such as profitability, can diverge. Less understood are some other properties of performance measures-and of performance measurement itself-which, taken together, suggest a need to rethink how firms assess their performance. The basic facts are these: The number of performance measures has burgeoned in recent years, while the performance measurement industry has grown rapidly at the same time. Oftentimes, however, performance measures still look like random numbers: For a given firm one measure may be high, another low, and a third in between.
Consider first how performance measures have burgeoned. Some years ago, firms paid attention mainly to industry-specific cost and output measures. Examples include advertising lineage and circulation for newspapers, cents per ton-mile for railroads, units delivered and market share in the automotive industry, gross sales volume in retailing, and crude measures of profitability-for example, profit as a percentage of sales. These measures have been largely retained over the years, but new ones have been added over time in response to changing conditions. ROA and ROI measures emerged in the 1920s, when the largest corporations adopted divisionalized structures and needed more precise ways to allocate capital internally. Measures more sensitive to shareholders’ interests, earnings per share and return on equity, followed in the 1950s and 1960s. Since the 1980s, considerable attention has been paid to cash flow measures. Indeed, some experts, such as Alfred Rappaport, argue that discounted cash flow is the best measure of corporate value and hence of shareholder value.
Meanwhile, growth in a performance measurement industry has paralleled the burgeoning of measurement tools. The Bureau of Labor Statistics reports that from 1983 to 1989, the number of accountants and auditors in the
Growth in performance measures and in the performance measurement industry is not by itself puzzling. Indeed, reasonable explanations can be found for both of these developments. What is puzzling, however, is that this growth has not produced better measures of performance, that is, measures that are better in a statistical sense. If there were an underlying property of firms called performance-which most economists assume but have never tested-then we would expect various performance measures, although flawed, to be correlated with one another and hence to gauge this underlying property. What we find, instead, is that most measures behave like share prices and profitability: They sometimes correspond, sometimes diverge, and occasionally diverge sharply. To be sure, weak or negative correspondences among performance measures are usually attributed to strategic choices made by managers. For example, market share can be increased at the expense of profits. Quality improvements can also erode short-term profitability. In fact, quality guru Joseph Juran writes that TQM can be expected to increase costs about ten percent in the short run. But while strategic choices may explain the discrepancies across performance measures, they do not account for the rapid growth in performance measures and the performance measurement industry.
One partial explanation for the puzzling properties of performance lies in the natural running down or deterioration of all performance measures over time, which requires new and different measures to be invented almost continuously. Running down occurs, in part, due to learning and adaptation: Given enough time, people discover ways to meet performance standards without changing their actual performance. Just as teachers are sometimes inclined to “teach to test” in order to demonstrate both students’ competence and their own, managers are sometimes tempted to engage in bookkeeping legerdemain in order to meet year-end goals.
The consequence of this tendency for performance measures to run down is diminished variation in performance outcomes and diminished confidence in existing performance measures-and a hue and cry for new and better measures. Changes occurring in the environment can also cause existing performance measures to run down. The recent experience of business schools underscores that tendency. For years, business schools have measured and rewarded research, and their research capabilities have grown substantially. Demands for better teaching now abound, and teaching performance is being measured more carefully. Given that teaching and research performance are not perfectly correlated-if they were, there would be no teaching problem-there are now two measures of performance rather than one. And there is also some confusion as to which standard accurately gauges performance.
A second partial cause of the puzzling properties of performance is in selection processes that preserve some firms and cause others to fail. Selection processes ultimately work against companies doing only one thing well: We know from research on small- and medium-sized organizations that the survival chances of highly specialized firms are diminished compared to companies retaining some slack and diversification. What we see in the
However, a problem potentially posed for financially driven companies such as GE is that the consolidation process never ends, since stringent profitability objectives force them continually to specialize (GE has sold most of its consumer electronic product lines) and consequently to downsize (GE has shed a third of its employees in the last ten years). Specialization almost always yields short-term savings, since attention is focused on a few objectives, and conflict over objectives is minimized.
Nonetheless, if our research findings hold for companies the size of GE, specialization can pose serious risks to long-term survival. It may be that firms whose shareholders are tolerant of slack and of apparently inconsistent performance outcomes are more likely to survive dramatic environmental jolts than firms applying a single performance yardstick rigorously. While the relationship of the variety and consistency of performance measures to firm survival is not known at this time, my own research shows that rapidly growing business units exhibit significantly greater variability in performance outcomes than stable or declining business units. Interestingly, the association between diversity of performance outcomes and business unit growth is stronger in mature industries than in industries at the early stages of the product life cycle.
A further explanation for the growth of performance measures lies in the search for measures of ever-increasing comparability. As long as most performance measures described costs and outputs, they remained industry specific. As a result, that rendered it difficult to combine different kinds of businesses in one corporation.
Meanwhile, it should be kept in mind that cost and output measures have little comparability across different businesses. Returns measures are comparable across different businesses within a firm but may be incomparable across firms because of differences in debt structures and accounting conventions, while cash flow measures are believed to be comparable across firms that are diversified internally. We know very little about the tradeoffs involved when attention shifts from industry-specific cost and output measures to accounting and financial measures having greater comparability across different businesses but lacking specificity. But the recent experience of
Which explanation best accounts for the puzzling properties of performance measures? Is it the natural running down of existing measures, selection processes favoring firms using multiple measures, or the search for measures of ever-increasing comparability? The answer may be less important than what these diverse explanations imply. Simply put: No firm can be managed according to a single performance standard for long. Moreover, new performance measures are emerging constantly and, as a consequence, some messiness and inconsistency in performance outcomes is not only to be expected but is also to be desired.
The question, then, is not whether current performance standards will change but rather how they will change. Dramatic change is imminent, partly because of the so-called quality movement. The movement is perhaps the most important development in
The quality movement is not oblivious to financial outcomes, but it takes as a matter of faith that better processes and hence better products and services will yield better profits-in the long run. Inevitably, firm-wide quality measures will emerge, perhaps along the lines of criteria for the Malcolm Baldrige National Quality Award. But these measures will not be perfectly correlated with short-term accounting and financial performance outcomes-somewhat to the dismay of managers driven to maximize shareholder outcomes but less attentive to production and to meeting customer requirements.
Marshall W. Meyer is Anheuser-Busch term professor of management and sociology at the