The economic analysis underlying corporate decision making: what economists do when confronted with business realities—and how they might improve

Business Economics, July, 2004 by Hugh Schwartz

Improving Heuristics and Including Them in the Guidelines for Analyzing Decision Making: The Final Recommendations

Many respondents commented that the texts on managerial economics and engineering economics are not helpful in resolving many problems they confront. Those texts will not become much more useful until they deal with situations in which optimization would be too costly or simply unobtainable, but in which profits can be increased by turning to second-best rules of thumb and taking account of their biases. Added to the need to alter managerial economics texts is the plea of one well-positioned industrial economist that the economics profession train people who can explain key concepts to non-economists within their enterprises. Ph.D. programs do not deal with how to communicate economic analysis to non-economists, nor do the economics courses in many MBA programs. Another respondent added that major departments of economics train students for academic or government positions (and banks, I would add), not positions in industrial corporations.

The variation in heuristics and their biases is so influenced by the specific contexts and the concerns of confidentiality are so great that at this stage at least, the most useful work will have to be done within the enterprises themselves. This may be achieved if the number of economists used in industry increases; if consulting firms employ suggestions of the type found in this report; and, perhaps most important, if this kind of material comes to be included in the micro and managerial economics texts that so many future executives are exposed to in business school programs. In the meantime, academic economists might undertake economic laboratory experiments that involve businessmen, not just students. Experiments might be designed that lack some of the components necessary for optimizing calculations, as is common in many business situations. Although the heuristics confirmed in laboratory experiments can contribute to a rational process of decision-making--it remains unclear to what extent laboratory experiments capture the decision-making behavior that can be expected in certain types of real life situations. It would also be desirable to study the behavior of business persons who participate in experiments when they are outside the laboratory, in order to determine how their performance in games maps onto actual behavior.

Business economists make assessments of heuristics at present, but they do so informally. Exposure to written and oral material on bounded rationality and behavioral economics facilitates recognition of the heuristics and routines used, along with their adequacy in given circumstances. It also helps the economists distinguish between a) recognition of those adaptive techniques as elements of a rational process of decision making and b) realization of the degree to which they represent deviations from rational decision-making. What I am suggesting could be incorporated into the preliminary evaluations of investments that some corporations undertake a year or more after commitment. At present, many corporate officials do not regard those exercises as useful. If guidelines for the categories of information that should be noted during the course of making the decisions and subsequently implementing them were drawn up by an internal consulting group in which economists participated, then the value of those evaluations might be increased and their worth might be better appreciated. In that case, more could be said about the specific heuristics employed, their suitability given the context, and the biases detected. It is true that those evaluations do not always lead to changes, even when the ex post audits focus on what needs to be done to alter the process--the management system--rather than on the specifics of a given project. Sometimes the findings of those ex post analyses simply do not register with a CEO, who may have his/her own approach to decision-making, perhaps imperfectly understood by other participants in the evaluation and, for that reason, not adequately taken into account in the evaluation.