CHAPTER IX PLACES CONSIDERABLE EMPHASIS upon the difficulty of measuring results in public organizations, and argues that it is easier to do so in business organizations. The latter have profits as the “bottom line,” and accountants know how to measure those—at least to a first approximation. But in spite of this, the measurement of results in business firms faces difficulties not unlike (although perhaps less severe than) those that public and non-profit organizations encounter.
First, there are difficulties with respect to tradeoffs between short-run and long-run profits (reflected, for example, in the interest rates assumed in estimating the present value of future income), and there are large ambiguities in such accounting items as “goodwill” and in the value of fixed assets that have been purchased (e.g., through mergers).
Second, even more severe problems are encountered when companies divisionalize with the aim of providing each division with a profit- and-loss statement and requiring it to make a profit. Then the pricing of transactions between divisions becomes a serious problem unless (and this is rarely the case) divisions are expected to deal with each other at arm’s length and to exhibit no preference for dealing with other divisions of the company rather than using outside vendors. Unless there are genuine competitive markets constraining the transactions among divisions, some administrative pricing procedure must be imposed in order to determine division incomes, expenses, and profits.
Third, there are many departments in any company (accounting, personnel, law, planning, research and development, advertising, etc.) whose contribution to profits is indirect, deriving from the services they provide to the “line” departments. In many if not most cases, it is not feasible to create internal markets for these services, and in any event, such markets would not be competitive and would be inadequate mechanisms for setting prices.
1. Long Run and Short Run
There has been a good deal of public discussion in recent years of reasons why company CEOs are tempted to emphasize short-term profits at the expense of the company’s future prospects. Given the relatively brief average tenure of executives at top corporate levels, when a large part of executive compensation takes the form of bonuses based upon annual profits, horizons of five years or even two years may have more weight in decisions than longer-run horizons. This is equivalent to using a very high interest rate to evaluate investment opportunities. (The same temptation is present for the occupants of positions in public organizations. Survival in any position may be facilitated, for a time, by postponing present problems to the future.)
All might be well if it were easy, or even possible, for stockholders to estimate the tradeoffs between present and future that are governing company decisions, but the major uncertainties that always control the future (and which are more easily assessed by those in management than by owners who are outsiders) make it exceedingly difficult to arrive at such estimates. Moreover, uncertainties about the future have a tendency to convert themselves into rather spurious “certainties” on the company’s books, where a definite value must be assigned to such items as “goodwill,” the value of factories and machinery and other tangible assets. For these reasons, stockholders’ equity on a balance sheet is seldom an even approximate estimate of discounted future earnings.
As no one, to the best of my knowledge, has proposed an adequate solution to this problem, I simply record it here as one of the obstacles to the effective use of the efficiency criterion in decision-making in private corporations as well as public agencies.
2. Divisional Profit-and-Loss Statements
How well outcomes can be evaluated in profit-and-loss terms (or in any other terms, for that matter) for subdivisions of an organization depends on the degree of mutual independence of the components. If the company is, in fact, a holding company, performing little more than an investment banking fimc- tion for its subsidiary divisions, then the divisional P&L statements can be as informative, or uninformative, as the statements for whole corporations.
A major consideration is whether the divisions are allowed to, and expected to, deal with each other at arm’s length. One aspect of this, as suggested above, is the right to choose between outside and inside suppliers on a basis of price and quality, without any special preference for inside purchases.
For transactions where the arms-length rule does not apply, the internal supplier and user will constitute a single buyer facing a single seller, and some procedure other than simple buying and selling in a competitive market will have to be provided to fix the “fair” price of transactions. Of course, if there are competitive external markets, the prices in these markets can be used as guidelines for the negotiated internal prices even if outside procurement is not permitted. But this is quite different from leaving transactions to the free play of markets.
There is a great deal to be said, of course, f or letting each tub rest on its own bottom. The divisional P&.L statement is an attractive device for enforcing efficiency, but only when the circumstances permit interdivi- sional transactions to be priced at reasonable approximations to prices in competitive markets. In using this device, however, one must remember that there is no magic in it that creates independence among units when technology, common marketing organizations, or other circumstances actually enforce a high level of interdependence among them.
3. Evaluating Intermediate Outputs
Perhaps the greatest difficulty in applying the efficiency criterion lies in evaluating those outputs of activity that are not final products. The problem is serious enough in constructing cost accounting systems for factories in such a way as to assign costs correctly to their causes. The problem becomes many times enlarged in assigning values to work that contributes mainly to the organization’s decision-making process. This includes, of course, all managerial activities, and especially those that feed only indirectly into production: research and development, legal services, advertising, and accounting being prime examples. In this domain, the difficulties that companies face in measuring the value of output are no less severe than those faced by government organizations.
The problem is illustrated by the rash of “downsizings,” and especially reductions in white-collar and middle-management personnel, that have been occurring in American corporations since the early 1990s. In some cases downsizing was a reaction to losses in sales volume, but in other cases, companies apparently decided that they could continue to operate at current levels with substantially fewer employees. If they were correct in those judgments, then their judgments about the required workf orce prior to downsizing must have been wrong.
Whenever a job is eliminated whose product has no immediate effect upon output or sales effort, short-run profits, as recorded in financial statements, can be expected to increase. What the effect will be on long- run profits is more problematic, and the answer will not be known for some time, if it is ever known. The arguments about staffing policies that take place in companies are no different from the arguments that take place in government agencies; they center upon the reality or unreality of the contributions that indirect activities make to the achievement of organizational goals over the more or less long run.
4. Evaluating “Quality”
We can divide into two (or sometimes more) parts the task of evaluating activities whose outcomes contribute only indirectly to the final product. First, we can usually find ways of evaluating the quality of the activity itself. Then we undertake the more difficult task of evaluating whether the activity, even if carried on at a high level of quality, is worthwhile in terms of final goals. As what may be a rather extreme example, consider the problem of evaluating a research professor at a research university—leaving aside for the moment his or her teaching contribution. The case would not be significantly different if we were evaluating a researcher in an industrial R&D lab.
Research in a university is a public activity. That is, the work is not complete until it has been evaluated and placed in the public record (published in a refereed journal). Then, it is evaluated again by those who do or don’t find it useful for their own subsequent research. It is evaluated also during the interactions that the researcher has with colleagues in the laboratory, in seminars, and at professional meetings. The evaluators are themselves knowledgeable, some to a very high degree, about the substance of the work. As a result, there is usually considerable consensus in any given research domain as to the quality of the output of any particular researcher, and it is not usually too difficult to arrive at a rough rank order of researchers in such a domain.
In scientific circles some cynicism is often expressed about weighing and counting pages of research output, and readers will be f amiliar with the phrase publish or perish. I do not want to make the evaluation process sound more valid than it is; but it is clearly far more veridical than weighing and counting. Its validity stems directly from the opportunities that professional peers have to observe both the process of research and its outcomes.
To the extent that the work is visible to, and thereby can be judged by peers, it affords the same opportunity for evaluation as occurs in research. People in organizations make judgments, therefore, about who is an effective or an ineffective manager, and about the level of effectiveness. Like judgments about research, these organizational judgments are fallible, but they are not random.
But both in the case of research and in the case of management what is being assessed is largely the short-run rather than the long-run value of the work. If we ask of a researcher’s output what long-run impact it will have upon the science or its application, that question will be harder to answer. If we ask of the manager not merely whether he goes about his work in an effective way but whether his important decisions are generally correct, then we are back to the task of tracing out chains of causal relations, many of them still hidden in the future. What is distinctive about downsizing, as contrasted with other dismissals of employees, is that it is based, not on a judgment that the work is being done badly, but on a judgment that, even if it is done well, it is not making an adequate contribution to the organization’s goals. As we have just seen, this is an exceedingly difficult judgment to make on an objective basis.
5. Competing Criteria of Evaluation
A very instructive example of the subdeties of evaluation, when it tries to go beyond direct measurement of the quality of activities, arose in the malaria control activities of the U.S. government just before, during, and after World War II.22 Malaria had long been a serious public health problem in many parts of the South, but its incidence decreased very rapidly (for reasons that were largely unknown) in the immediate prewar years. During the war, many American soldiers were exposed to malaria abroad, and many contracted it, leading to the fear that new epidemics might break out when they returned to this country. Public health agencies used a variety of programs to deal with the disease, one of the most extensive being large-scale insecticide (principally DDT) attacks on the mosquitoes that transmit it.
There were, therefore, two kinds of measures of the effectiveness of malaria control: numbers of reported cases (and of deaths), and sizes of mosquito populations. Unfortunately, diagnosis and reporting of cases and deaths was very unreliable, for definitive blood smear tests were not widely used during much of this period. Nevertheless, the evidence was strong, although not always accepted by the scientists, that by 1942 the agency was combating a disease that was nearly extinct in the United States. However, buoyed by the uncertainties of the morbidity and mortality statistics and by the prospects of reinfection of mosquitoes by returning veterans, the agency turned from morbidity and mortality statistics to the statistics on mosquito populations as the basis for effort evaluation. They could then show that the populations were (potentially) dangerously large, and that DDT spraying led to major reductions in them. During a period during which the agency spent approximately $50 million, the question of the relation of the activities of the agency to the continuing absence of malaria remained essentially unanswered.
It is not clear whether better statistics in this case would have made it easier to arrive at correct policy decisions. The most important information that was missing was information about the causes for the apparently unmotivated but nearly complete disappearance of the disease at the beginning of the 1940s, and information about the probabilities that a new epidemic could be started by the return of infected veterans.
Is comparable uncertainty common in business decision-making? Consider a multinational company (Sea Containers, Inc.) with one division in the container ship business, another in passenger ferries on several European ferry crossings, a third in luxury hotels on several continents, and a fourth operating railroad trains (e.g., the Orient Express). Decisions have to be made frequently about the purchase and disposal of ships and containers and the acquisition and termination of ferry routes. How should the long-term effects of the English Channel tunnel upon the Channel ferry business be estimated? What fluctuations can be anticipated in the container ship business, and what are the likely near-term strategies of competitors with respect to the purchase of new container capacity? What measures can be used to evaluate the effectiveness of the company’s managers? The fact that the company has a “bottom line,” and that its short-run return on investment can be measured, does not in fact make its problems of evaluation very different from those of a government agency.
Source: Simon Herbert A. (1997), Administrative Behavior, Free Press; Subsequent edition.