The 1990s are shaping up to be a decade of dramatic changes. The recent shake-ups at IBM and General Motors are ominous signs that “business as usual” will be unusual in the coming years. The current changes require more than the usual reorganization or restructuring — they require a redesign of our companies and economic systems. Jay Forrester lays out the new challenges we face as we enter our latest frontier.

The continued search for better understanding of social and economic systems represents the next great frontier in human development. Frontiers of the past have included creating the written literatures, exploring the geographical limits of earth and space, and penetrating the mysteries of physical science. Those are no longer frontiers; they have become a part of everyday activity. By contrast, insights into behavior of social systems have not advanced in step with our understanding of the natural world. To quote B. F. Skinner:

“Twenty-five hundred years ago it might have been said that man understood himself as well as any other part of his world…. Today he is the thing he understands least. Physics and biology have come a long way, but there has been no comparable development of anything like a science of human behavior… Aristotle could not have understood a page of modern physics or biology, but Socrates and his friends would have little trouble in following most current discussions of human affairs.”

Consider the contrast between great advances during the last century in understanding technology and the relative lack of progress in understanding economic and managerial systems. Why such a difference? Why has technology advanced so rapidly while social systems continue to exhibit the same kinds of misbehavior decade after decade? I believe the answer lies in failing to recognize that countries and corporations are indeed systems. There is an unwillingness to accept the idea that families, corporations, and governments belong to the same general class of dynamic structures as do chemical refineries and autopilots for aircraft. To admit the existence of a social system is to admit that the relationships between its parts have a strong influence over individual human behavior.

The idea of a social system implies sources of behavior beyond that of the individual people within the system. Something about the structure of a system determines what happens beyond the sum of individual objectives and actions. In other words, the concept of a system implies that people are not entirely free agents but are substantially responsive to their surroundings.

To put the matter even more bluntly, if human systems are indeed systems, it implies that people are at least partly cogs in a social and economic machine. People play their roles within the totality of the whole system, and they respond in a significantly predictable way to forces brought to bear on them by other parts of the system. This is contrary to our cherished illusion that people freely make their individual decisions. I suggest that the constraints implied by the existence of systems are true in real life. As an example, we see the dominance of the political system over the individual in the evolution of the Federal budget deficit. Every presidential candidate since 1970 has campaigned with the promise to reduce the federal deficit. But the deficit has on the average doubled every four years. The social forces, rather than the president, have been controlling the outcome. How to harness those social forces has not been effectively addressed.

Designing Managerial and Social Systems

In designing an engineering system such as a chemical plant, engineers realize that the dynamic behavior is complicated and that the design cannot be successfully based only on rules of thumb and experience. There would be extensive studies of the stability and dynamic behavior of the chemical processes and their control. Computer models would be built to simulate behavior before construction of even a pilot plant. Then, if the plant were of a new type, a small pilot plant would be built to test the processes and their control.

But observe how differently social systems are designed. Although political, economic, and managerial systems are far more complex than engineering systems, only intuition and debate have ordinarily been used in building social systems. We change laws, organizational forms, policies, and personnel practices on the basis of impressions and committee meetings, usually without any dynamic analysis adequate to prevent unexpected consequences.

“Designing” social systems or corporations may seem mechanistic or authoritarian. But all governmental laws and regulations, all corporate policies that are established, all computer systems that are installed, and all organization charts that are drawn up constitute partial designs of social systems. Such redesigns are then tested experimentally on the organization as a whole without dynamic modeling of the long-term effects and without first running small-scale pilot experiments. For example, bank deregulation and the wave of corporate mergers in the 1980s constituted major redesigns of our economy with inadequate prior consideration for the results. All systems within which we live have been designed. The shortcomings of those systems result from defective design, just as the shortcomings of a power plant result from inappropriate design.

Effects of Feedback Structure

The feedback structure of an organization can dominate decision making far beyond the realization of people in that system. By a feedback structure, I mean a setting where existing conditions lead to decisions that cause changes in the surrounding conditions, that influence later decisions. That is the setting in which all our actions take place.

We do not live in a unidirectional world in which a problem leads to an action that leads to a solution. Most discussions, whether in board meetings or cocktail parties, imply a structure which suggests that the world is unidirectional, that the problem is static and we need only act to achieve a desired result (see “Open-Loop Impression of the World”).

Instead, we live in an ongoing circular environment in which each action is based on current conditions, such actions affect conditions, and the changed conditions become the basis for future action (see “Closed-Loop Structure of the World”). There is no beginning or end to the process. People are interconnected. Through long, cascaded chains of action, each person is continually reacting to the echo of that person’s past actions as well as to the past actions of others.

In general, social systems carry a set of common characteristics:

• Most difficulties are internally caused, even though there is an overwhelming and misleading tendency to blame troubles on outside forces.
• The actions that people know they are taking, usually in the belief that the actions are a solution to difficulties, are often the cause of the problems being experienced.
• The very nature of the dynamic feedback structure of a social system tends to mislead people into taking ineffective and even counter-productive action.
• People are sufficiently clear and correct about the reasons for local decision making — they know what information is available and how that information is used in deciding on action. But people often do not understand correctly what overall behavior will result from the complex interconnections of known local actions.

Jay W. Forrester, Professor Emeritus at the Massachusetts Institute of Technology and former director of the MIT System Dynamics Group, is the founder of the field of system dynamics. Since his retirement in 1989, he has been working toward bringing system dynamics into K through 12th grade schools as the basis for a new kind of education.

Analyzing Social Systems

A systems analysis of a company draws on knowledge about how structure and policy relate to behavior. Information comes primarily from interviewing people in the company about how they make decisions at their individual operating points. Statements describing the basis for decisions are the rules or policies governing action (“policies” meaning all the reasons for action, not just formal written policy). There is substantial consistency throughout the organization as to the actual operational policies that are guiding decisions. Furthermore, the policies are justified in terms of how those policies are expected to correct the great difficulty that the company is experiencing.

Up to this point, the study of such a company follows the case-study approach to management education. That is, a comprehensive examination of all related parts of the company is made in the context of the problem that is to be solved. But if left at this point, the weakness of the case-study method would dominate the outcome. A descriptive model of the company would have been assembled, but the human mind is not able to deal with the inherent dynamic complexity of such a situation.

Such a descriptive model is equivalent to a high-order nonlinear differential equation. No scientist or mathematician can solve such a system mentally. Only computer simulation methods are capable of revealing the behavior implicit in the structure that can be built from knowledge about the many local decision-making individuals and how they are connected.

After obtaining a description of the important policies, information flows, and interconnections in a company, the next step is to translate that description into a computer simulation model. A simulation model does not involve complicated mathematics but instead is a language translation from the original description to computer instructions. Such a model allows the computer to act out the roles of each decision point in the model and feed the results to other connected decision points to become the basis for the next round of decisions. In other words, a laboratory replica of the company then exists in the computer where one can observe the behavioral consequences of the policies that had been described in the interviews — policies that are intended to solve the company’s problem.

To the surprise of those unfamiliar with the devious nature of such dynamic systems, the computer model, based on policies known to people in the company, will usually generate the very difficulties that the company had been experiencing. In short, the policies that were believed to solve the problem are, instead, the cause of the problem. Such a situation creates a serious trap and often a downward spiral. If the policies being followed are believed to alleviate the problem, but, in hidden ways, are causing the problem, then, as the problem gets worse, pressures increase to apply still more strongly the very policies that are causing the problem.

Similar misjudgments lie behind much of the foreign encroachment on American markets in the 1980s. Foreign infiltration was initially blamed by American companies on lower foreign wages and lower product price. In response, domestic prices were reduced until there were insufficient profit margins to permit fixing the real difficulties, which were usually more in design and in quality of product and service than in price. As so often happens, the domestic failure to compete arose more from mismatched internal policies than from external forces.

A New Kind of Management Education

All of this points to the need for a new kind of management education, one that we might call “enterprise design.” It also suggests a new kind of manager for the future — the “enterprise designer.”

A fundamental difference exists between an enterprise operator and an enterprise designer. To illustrate, consider the two most important people in the successful operation of an airplane. One is the airplane designer and the other is the airplane pilot. The designer creates an airplane that the ordinary pilot can fly successfully. Is not the usual manager more a pilot than a designer? A manager is appointed to run an organization. Often there is no one who consciously and intentionally fills the role of organizational designer.

Organizations built by committee, by intuition, and by historical happenstance often work no better than would an airplane built by the same methods. Time after time one sees venture capital groups backing a new enterprise in which the combination of corporate policies, characteristics of products, and nature of the market are mismatched in a way that predetermines failure. Like a bad airplane design that no pilot can fly successfully, such badly designed corporations lie beyond the ability of real-life managers.

Management education, in all management schools, has tended to train operators of corporations. But there has been rather little academic attention to the design of corporations. The determination of corporate success and failure seldom arises from functional specialties alone, but grows out of the interactions of functional specialties with one another and with markets and competitors. The policies governing such interactions have not been adequately handled in management education. We need to deal with the way policies determine corporate stability and growth in an intellectual, challenging, quantitative, and effective way. Such management education leads to what I refer to as enterprise design. Such an education would build on three major innovations that have already occurred in this century: the case-study method of management education as pioneered by the Harvard Business School beginning around 1910; the development of theory and concepts related to dynamic behavior of feedback systems as first developed in engineering at the Bell Telephone Laboratories and MIT in the 1930s and 1940s; and computers that permit simulation modeling of systems that are too complex for mathematical analysis.

Bringing these three innovations together offers the potential for a major breakthrough in management education by adding a rigorous dynamic dimension to the rich policy and structural knowledge possessed by managers. The difference between present management schools and those of the future will be as great as the difference between a trade school that trains airplane pilots and a university engineering department that trains aircraft designers. Pilots continue to be needed, and so will operating managers. But just as successful aircraft are possible only through skilled designers, so in the future will competition create the necessity for enterprise designers who can reduce the number of design mistakes in the structure and policies of corporations.

In management there is a tendency to identify a weakness, then try to find ways to relieve the symptoms. But it would be more fundamental to insist on understanding why the objectives are not already being met. What is it in the design of a corporation that is inhibiting success? A frontal assault on the symptoms, while the underlying causes remain in place, almost always fails. Success will follow when the designs of corporations give greater emphasis to removing the causes of problems rather than to trying to counteract the symptoms.

Forthcoming as “System Dynamics and the Lessons of 35 Years,” in Kenyon B. De Greene (ed.) Systems-Based Approach to Policy-making (Norwell, MA: Kluwer Academic Publishers, 1993).

Jay W. Forrester, Professor Emeritus at the Massachusetts Institute of Technology and former director of the MIT System Dynamics Group, is the founder of the field of system dynamics. Since his retirement in 1989, he has been working toward bringing system dynamics into K through 12th grade schools as the basis for a new kind of education.

Further reading: Industrial Dynamics (Cambridge, MA: Productivity Press, 1961); Urban Dynamics (Cambridge, MA: Productivity Press, 1969) Collected Papers of Jay W. Forrester (Cambridge, MA: Productivity Press, 1975), World Dynamics (Cambridge, MA: Productivity Press, 1973).