Succeeding in economics.

Demsetz, Harold

The editors of American Economist invited me to discuss professional aspects of my life as an economist as part of an ongoing series of similar discussions by others. They thought readers might find this of interest and might even benefit from learning how contributors to this series go about their work. My intent here is simply to highlight some aspects of my career that have helped me progress in the profession. What has helped me may not be equally helpful to others, each of us being unique in experience, capability, and interest. Indeed, some conditions that have helped me along have not been of my own making. Each of us, for example, is born at a particular time in a particular place and to a particular set of parents. Although we can influence the views of our parents as we mature, we are powerless to affect time and place of birth. Yet, these do influence what we will do and how well we will do it.

The first full year of the Great Depression, 1930, is the date of my birth and the city of Chicago is the place. The decade of the 1930s, of course, began a long period of hard times for a great many families, including mine. Although hard times make success more difficult in ways that are well known, the 1930 birth date made success in academia easier. One advantage of this date is that I became a member of an age cohort group that was made smaller by the Depression. When I entered the academic marketplace in 1958 I faced fewer competitors than would have been the case if prior years been more prosperous. Another advantage was that the GI bill, which offset what would have been the impact of prior Depression years, was still channeling very large numbers of WWII and Korean War veterans into college as I sought my first professional teaching position. These consequences of my birth date put me in the academic marketplace when the entering supply of new PhDs was relatively small and the demand for their services was relatively large. Adjusting for quality of training, most graduating PhDs enjoyed an abundance of job offers that lasted for the better part of a decade.

Being born in a large city was also advantageous. The neighborhood of my youth was largely populated by immigrant families from Europe who, though not very knowledgeable about post high school educational institutions, were self-selected for their willingness and ability to deal with circumstances new to them. College presented such circumstances. The first year of my college education took place at an essentially free nearby public city college, one, as it turned out, attended by ample numbers of bright students and serviced by capable teachers who probably would have secured better positions if prior years had been better. The second year of my education was at the Chicago undergraduate division of the University of Illinois, where tuition at the time was about $100 per semester. The cost of going to college in a large city at that time was also kept low by being able to live at home with the family on Chicago's West Side.

A large city also offers opportunities to pay for an education through part time work; during summers I drove a Checker cab and during the school year I ushered evenings at the Chicago Civic Opera House. I studied engineering during these two years and, therefore, learned some mathematics. As part of the distribution requirements for the engineering degree, I took a course in economics. This was taught well by an instructor who was then working on his PhD at the University of Chicago.

I found economics interesting and enjoyable. The Opera House job helped round out my education in the arts as I pursued work in the engineering program. Financing an education this way has its downsides. Commuting took the better part of three hours a day and so did part time work. Being raised as a user of urban transport, I did not fully appreciate the weight of this burden until I continued my education at the Champaign-Urbana campus of the University of Illinois. After a brief tour with the U.S. Army, which made me eligible for G.I. bill aid, I went on to secure an MBA and PhD from Northwestern University.

I did not consciously engage in the long-run planning of my teaching-research career when I entered the profession, as I did not begin my career with any expectation or intent of making a "mark" in the profession; indeed, I saw college teaching as a comfortable job to which a three month vacation was attached. As things turned out, I became increasingly interested in economics as I read more of it and taught my classes. I began writing on topics I found especially interesting, and perhaps because of being raised on Chicago's West Side I experienced little apprehension about going public with my work. As things have turned out, I have never, I say never, taken even a month away from work. Economics turned out to be too interesting and idea production too stimulating. Not everyone can find a remunerative work that keeps a person busy and satisfied at the same time, but I did. It is difficult, I think, to make a confident judgment about this without first really working the tasks that a career sets before a person.

Although I have written on many topics, my writings are mainly contained within three categories of subject matter: (1) markets and firms; (2) property rights and externalities; and (3) financial markets and transaction costs. This degree of concentration has worked well for me. It offers a comfortable compromise between interesting targets of opportunity and continuity of knowledge accumulation.

Sources of potential topics to work on are many. Examples of where some of my ideas have come from may be given. One of my earliest publications arose from my studies at Northwestern University while working toward my PhD. I had been reading Edward Chamberlin's Theory of Monopolistic Competition (1933), a book that had become a mainstay of industrial organization theory during the 1950s. An aspect of the way Chamberlin developed his notion of excess capacity equilibrium puzzled me, and I kept mental note of this. The cost curve he used to describe his excess capacity theorem assumed implicitly--without his taking note of it--that promotional expenditures are held constant while a firm varies its output rate; otherwise how could the demand curve facing the firm remain in the fixed position given to it by Chamberlin? In contrast to this, all costs are allowed to vary optimally when discussing the cost curve of the perfectly competitive firm. Yet, despite the difference in the implicit definitions of these two cost curves, Chamberlin's notion of excess capacity was based on the supposed deduction that the equilibrium rate of output of the monopolistically competitive firm is less than that which would bring unit cost to its minimum if the market were perfectly competitive; true, but this minimum cost is associated with a very special cost curve, one which holds promotional cost constant. The incompatibility between the two cost curves made me wonder about the equilibrium of a monopolistically competitive firm if full consideration is given to the variation in optimal promotional expenditures that surely would take place as output rate varies. At a later time, but before I had completed my PhD work, I returned to this question. The result was publication of "The Nature of Equilibrium in Monopolistic Competition" (J. Pol. Econ. Feb. 1959).

The teaching experience is a second source. An essay titled "Purchasing Monopoly," in my Efficiency, Competition, and Public Policy (Oxford: Basil Blackwell Ltd. 1989), comes from a mental note I made while teaching graduate students at UCLA. The textbook treatment of the reward to monopoly is calculated as the difference between revenue received by a monopolist and cost incurred by the monopolist. In this calculation, cost is treated as expenditures made to produce the monopolist's profit maximizing rate of output, and the cost conditions that determine these expenditures are those that would define the supply curve of the competitive industry that would exist if their were no monopoly. This would be correct if monopolizing a market required no special expenditures, as if the monopoly were a gift of nature or of a costless conversion of a competitive industry. While discussing the case, I realized that a "free" monopoly would not ordinarily be the case. Rivals will need to be defeated or purchased, and would be entrants would need to be bought off. Monopolizing would at least require the payment of Ricardian rents to firms in the competitive industry that is being transformed into a monopoly. These monopolizing costs are not incurred by firms in a competitive market. I returned later to examine the issues raised by this realization. The essay that came from this shows, or course, that the return to monopolizing is much less than the return to the receiver of a free gift of monopoly, so the customary measure of monopoly profit exaggerates the true return to the monopolist who, in one way or another, must buy the right to exclude rivals. More than this, the essay also shows that the use of deadweight loss triangles to measure the social loss of monopoly yields a biased estimate.

A third idea source is conversation with colleagues. At a lunch shared with colleagues at the University of Chicago, discussion turned to a report about antitrust that had just been written by a committee headed by Phil Neal, then Dean of the University of Chicago Law School. One part of the report endorsed the market concentration doctrine. This well-known doctrine had guided thought about the relationship between market structure and pricing power when I was a young man, and to some extent it still does. An observation was made by one of my colleagues at this lunch. This was that domestic auto industry profits were comparatively high only because GM's profit was high. Neither Ford nor Chrysler, for example, earned exceptionally high profits. I began investigating whether this observation had generalizing power when I arrived at UCLA in 1971. In the absence of governmentally imposed barriers to entry, market concentration may be expected to arise if scale economies are important or if innovation has resulted in a superior product or way of producing an existing product. Competitive responses to these sources yield not just variation in market structure across industries but also variation in measured industry profit rates. Examined in more detail, market concentration as an outcome of efficient structure is consistent with a correlation between profit rate and market concentration if the correlation is calculated only for the larger and historically more successful firms in the markets being compared, but the efficient structure hypothesis would not lead to an expectation of this correlation for middle sized and smaller firms. The higher profits received by the larger, historically more successful firms is to be attributed to their lower costs or their superior products given that product price is determined by cost at the margin in an industry. In contrast to this explanation, a pure collusion explanation would lead one to expect a positive correlation between profit and market concentration for all size classes of firms; evidence for this is nowhere found in the data (see, Demsetz, H. "Industry Structure, Market Rivalry, and Public Policy," J. of Law & Econ. April 1973 and Demsetz, H. "Two Systems of Belief about Monopoly" in Goldschmid, Mann, and Weston (Industrial Concentration: The New Learning; Little, Brown, 1974).

This example is especially useful in revealing a characteristic of some of my work. I tend to push an argument or a model back to an earlier analytical step than is presently covered by it; rather than take market concentration as a given when evaluating data, I first sought to explain variation in market concentration. This tendency is also illustrated by my article "Toward a Theory of Property Rights" (AER. May 1967). While I was at Chicago, R. H. Coase, who arrived a year after I did, amazed the profession by demonstrating, in the context of the externality problem, that who does (and who does not) possess a fight to impose a cost on others is irrelevant to efficient resource allocation. (This result obtains if transaction cost is zero and if income effects on demands for goods involved in the interaction are unaffected by wealth redistribution.) Coase's work took the private property fight system as an "existing" given when making his point. I began to think about moving the problem to a prior step by seeking to understand the emergence of a private fight system from a preexisting communal right system. This resulted in a simple efficiency explanation for the rise of private rights (and, under specifiable conditions, also for the persistence of communal rights). I explained how changing conditions brought about by the development of the fur trade in the new world affected ownership of land arrangements among Native Americans on the American Northeast and failed to do in the American Southwest.

One more tendency of mine might be noted. The interesting problems seem me to be those that begin with observation of an unexpected practice, organization, or business policy. In attempting to make sense of these, I usually begin by seeking an explanation that reconciles the anomaly with efficient allocation of resources. In the Native American land ownership example discussed above, the problem was to explain why Native American's of the Northeast came to treat land differently from Native American's of the Southwest. I found my answer in the different habits of forest animals and animals of the great plains. Each land ownership arrangement seemed efficient in the context of the animal habits facing Native Americans in the two locations.

The work I choose to do is not heavily armored with math and econometrics. It is focused on empirical and policy problems and on the logic of the theory that bears on these. Cold logic, imagination, and exposition by way of words, simple geometry, and basic statistics are the tools on which I have mainly relied throughout most of my career. I do not feel fully in command of a problem or of a resolution of it until I can state both clearly in words and/or geometry. This gives my work a wider audience but probably also limits the nature of the problems on which I choose to work. Not many young economists adopt this working methodology today, and I am not sure it is suitable for an initial job search in today's market. Today, prospective employers seem to strongly emphasize demonstrable ability in econometrics and mathematics. I suspect that the large degree of specialization across fields of economics explains this, since it is now more difficult to have the knowledge of material in the many fields in which the renderer of judgment is not a specialist. Technical tooling offers a substitute, but heavy attention to these tools often hampers discovering the economist. Quality of economic thinking is more likely to be revealed by the way the underlying problem to which the tools are applied is conceived and analyzed. Economists in training seem to seek a body of data that is appropriate for the exploitation of technical tools rather than one which offers intrinsic economic interest. Not frequently, but still on occasion, a seminar is given by a young economist who has uncovered such a problem and is now busily engaged in a thorough investigation of it; this person has a good chance of becoming a successful economist.

I do regularly attend workshops in my field, and I recommend this as a method for keeping in touch with the profession, for learning from other attendees, and for exposing others to ideas that may come to your mind. A second method is to become a member of dissertation committees and eventually to become a chair of some of these. This is an effective way to learn one's own strengths and weaknesses as well as those of students. It is the most efficient way to learn about the intellectual history of the topic featured in the dissertation. Beyond this, graduate students whose dissertations you chair become well acquainted with your work, ideas, and methods and they carry these to other institutions.

One reason I came to UCLA from Chicago is to have the opportunity to interact with students in their dissertation writing stage. My position at Chicago was in the Graduate School of Business, a fine school with fine students and colleagues, but most of these students sought only the MBA degree. Those who went for the PhD wrote dissertations in accounting, marketing, management, and finance; few wrote in economics. Chicago, after all, had a world class Department of Economics to service students seeking to become economists. My position at UCLA, in contrast, was in the Economics Department, where I managed to supervise the dissertations of almost 80 students. Times are different now, I suppose; a few more students in business schools seek a PhD in something called business or managerial economics.

A position in a top business school has an important advantage. This is to encourage work on puzzles involving business and market arrangements and on public policies toward business. Economists in economics departments have a stronger tendency to look within economics itself for the problems they study. The issue is not which of these problem sets is better, but which mixture of them works best for you. My personal development has benefitted much from practicing my discipline not only in a fine business school and fine economics department but also from teaching in fine law schools. The topics on which I have written reflect this three-fold institutional involvement. I began work in two economics departments, continued in Chicago's business and law schools, and then returned to economics. Along the way, I did my share of consulting on antitrust cases and institution building. Engaged thus, it is not easy to become bored. It goes without saying that one must publish to succeed in economics within academic institutions. Quality teaching performance and involvement in academic institutional problems count, but these provide no answer to the question "Is he or she a fine economist?" Quality of writings supplies this information.

One may write for different audiences. An economist who is good in writing for one audience may not be equally good for another. I recall George Stigler asking me more than once if I thought Milton Friedman will be most remembered for his scientific work (say, on the consumption function and monetary theory) or for his reformist policy work (as in his Capitalism and Freedom)? The answer George sought was "Milton's scientific work," but the answer, I think, depends on the audience being asked the question. The Nobel Prize Committee, representing the economics profession, stressed Friedman's professional work, and it may well be that future economists will do so also. But I think a broader "committee" of influential intellectuals might well value his reformist work more. To succeed professionally in your own terms, know your audience. Not many of us are as capable as Friedman was in serving two such different groups so well.

It is extremely important to become associated with the best economists available to you. This usually means appointment at the best universities possible. People think about puzzle resolution in different ways. Only by reading what others write and by engaging in discussion with them can you begin to discover different ways of looking at a problem. You also learn to disagree with your colleagues and to bear their disagreement with you--all in proper spirit. It does not do you or them much good if you to keep your thoughts to yourself. For a month or two I sat quietly, saying nothing, in Chicago's Industrial Organization Workshop. One day, George Stigler, the creator and director of the workshop, leaned over to me and whispered "You owe it to others to make your thoughts known." Except for those few months, I always do. Association with the best also broadens the invitations you will receive to make your thoughts available at other institutions.

I have been able to interact with some very fine economists while learning my trade. Perhaps this reflected an ample share of luck. Perhaps it reflected my ability and willingness to engage. However it came about, this has proved extremely important to my professional progress. Surely other "things" also matter, but I strongly advise young economists to increase the weight they give to associating with the best possible economists when seeking a position.

My work habits are not atypical. I do not often work into the evening hours, preferring instead to do my professional work during daytime in my office at the university. I come to my office almost every weekday, and I hardly ever take off more than a week or two during the calendar year. This remains my pattern even though I am now several years past official retirement. I see myself as a "long distance runner" when it comes to work; I read and writing regularly during daytime hours through most of the year. Once a work project is completed I quickly search for another if I do not already have one in inventory. Weekends are spent at home with my family, sometimes tending to repair work but always enjoying their company. On several evenings of each week my wife and I take three-mile hill walks. On the occasion of these walks we sometimes discuss the new work I am doing. Rita is quite good at understanding what I am saying and often senses problems that I have not dealt with. Having a great partner in life makes it much easier to succeed professionally. I have no personal experience with and can only imagine the considerable difficulties that face colleagues whose minds are occupied with problems at home or who do not look forward to closing the office door after a good day's work and returning home.

Harold Demsetz, Professor Emeritus, UCLA