Cowles Commission for Research in Economics

REPORT FOR PERIOD JULY 1, 1952 – JUNE 30, 1954

The University of Chicago

PURPOSE

The Cowles Commission for Research in Economics has as its purpose the conduct and encouragement of research in economics, finance, commerce, industry, and technology, including problems of the organization of these activities and of society in general. Its approach is to encourage and extend the use of logical, mathematical, and statistical methods of analysis. It seeks to accomplish its purpose through research and instruction, through publication, and through other programs directed toward increasing the human resources devoted to such research.

The Commission is a not-for-profit corporation, founded in Colorado in 1932 and, since 1939, chartered under the laws of the State of Illinois. Its governing bodies are its faculty, an executive committee, and a board of trustees.

Although it is an independent research organization with members of its staff in other research centers both in the United States and abroad, the Commission is affiliated in academic matters with the University of Chicago as a component of the Division of the Social Sciences. The Commission is also affiliated with the Econometric Society, an international society for the advancement of economic theory in its relation to statistics and mathematics.

INTRODUCTION

The role of academic research in a free society is the creation and communication of new knowledge. The search for new truth is governed by the demands of free inquiry: free choice of topic, free choice of method, and free development of conclusions. Communication, to be effective, must relate new knowledge to old, must trace new implications for social means and ends, and must describe the new relations in terms understandable first to the specialist and ultimately to the interested layman. Both freedom of inquiry and effective communication are essential to the development and ultimate fulfillment of the free society.

The Cowles Commission for Research in Economics presents here an account of its research efforts for the two-year period, July 1, 1952–June 30, 1954. The work described is the product of a community of scholars pursuing their interrelated research interests in an academic environment which is both stimulating and free.

Cowles Commission staff members are chosen for their interest and competence in the rigorous development of economic theory and measurement. They have at their disposal the technical and general literature of the University of Chicago library system as well as the specialized resources of the Commission's own library. They have free access to the knowledge and critical judgment of colleagues through regular staff meetings and through the distribution of preliminary discussion papers which describe their work in its early stages. Final research results are carried to publication in the form of monographs or journal articles. Reprints of journal articles are frequently given wide distribution as Cowles Commission New Series Papers.

Thus through publication and broad circulation of research results an attempt is made to achieve effective communication with other researchers. However, the technical nature of the Commission's research methods and results has sometimes made it difficult for nonmathematical economists and interested laymen to appraise or follow what is being done and why. For this reason the character of Cowles Commission general research reports, such as this one, has been changed in recent years(+) toward a form of exposition which, it is hoped, will be understandable to a greater number of readers. For those with research backgrounds and interests, the listings of papers, publications, and addresses following the research report provide a key to the technical scope of the Cowles Commission's work.

The research described in the present report is financed through general support by Alfred Cowles and other members of the Cowles family, through time devoted to research by members of the faculty of the University of Chicago, and through grants or contracts for specific research projects. The study of futures markets has been supported by a grant from the Rockefeller Foundation. The studies of transportation systems have been made under a research contract with the RAND Corporation on the theory of resources allocation. The work on organization theory and the studies in the theory of competitive markets are conducted under a research contract, on decision-making under uncertainty, with the Office of Naval Research.

By developing financial support for the establishment and maintenance of a body of able scholars who are free to pursue their own research interests in their own way, the Cowles Commission promotes freedom of inquiry in the social sciences and thereby fulfils part of its responsibility as an academic research organization. By extensive dissemination of research results and by striving for simplicity and clarity in its annual or biennial research reports, it is attempting to broaden understanding of the economic problems with which it deals. Current readers and future years will determine whether or not this has been done effectively.

REPORT ON RESEARCH ACTIVITIES

The educated citizen looks upon economics as a rather complicated but probably dull subject. He is likely to admit that the development and the spread of economic knowledge are important for the welfare of society. He may even look with favor on the slow and halting process of accumulation of knowledge called economic research. But rarely does he take a personal interest in following the progress of economics. He is deterred equally by the abstruseness of theoretical writing — if he ever sees it — and by the use economists make of masses of unappealing statistical data on production, trade, consumption, inventories, and prices. If it is further intimated that mathematics — a subject he generally avoids — is used in economics to untie the knots and follow the threads of causation, then he is likely to leave the subject alone. For the satisfaction of his urge to know he may turn to history, or to psychology, or to any other field closely related to human experience. Or he may study the victories of mind over matter in technology, over nature in medicine.

It is not contested that the vindication of the effort that goes into research on such a practical subject as economics must lie in increased ability of mankind to make the best of its circumstances, resources, and endowment. However, the appraisal of what economists are contributing in this regard would be helped if their subject could be made to come to life more fully to the interested citizen. Somehow economists have so far not succeeded in putting across the fact that, back of every statistical series in the record of economic history, there is a whole category of human decisions. These may be decisions taken in the organization of production or trade, in the pursuit of gratification, as precaution for contingencies, or in bargaining for gain or advantage. There is a human decision around every corner in economic analysis. It is this aspect that makes economics a complicated, it is true, but an equally fascinating subject.

One source of fascination is that here, as in psychology, man studies an important part of himself. In addition, there is perhaps a fascination peculiar to economics in tracing the effects of interaction between different individuals, or groups of individuals, each with his own motivations or objectives. While the objectives of different individuals are partly conflicting, no one controls all the variables. One man may offer a commodity at a given price, but another will decide how much to buy at that price. One man may decide what level of inventory to aim for in setting the rate of production, but many other people by their purchases decide indirectly whether the inventory will actually be at that level or higher or lower. A union may call a strike or threaten to do so. An employer may decide to have it out or may forestall interruption of production by concessions. Since neither party to a contract can dictate the other party's behavior, the outcome reflects the bargaining power — that is, alternatives available — on the two sides. Generally, the outcome of interaction between economic decision-makers comes to less for each party than he would like. Sometimes, it is a best achievable compromise — in some all-around sense — that is permitted by the circumstances. But always it is a fascinating subject for analysis, full of surprises.

1. Studies of Markets

Markets are the stages on which the interplay of economic decision-makers is acted out. In the Cowles Commission's research program during the two-year period, July 1, 1952–June 30, 1954, a great deal of attention was given to the empirical and theoretical study of markets. Two empirical studies, one of the markets for future delivery and one of the market for livestock and livestock products, and one theoretical study of the action of competitive markets were mentioned in our previous ("Twenty Year") report and can now be described in greater detail.(+)

Both the empirical and the theoretical studies follow the method of "model construction": important variables are defined and, in the empirical studies, measured statistically. The behavior of the various parties in the market is described by a set of mathematical relationships between these variables, called a "model." In the theoretical studies the implications of such relationships are traced. In empirical studies the relationships are also estimated numerically or graphically to give greater quantitative precision to the conclusions.

The study of futures markets in corn, cotton, and wheat conducted by Houthakker, assisted by Telser, is nearing completion and is expected to be published in monograph form. One of the main purposes of this study is to throw light on the economics of behavior under uncertainty through the investigation of markets that have been developed to deal with that problem in practice. It is hoped, however, that the results will also be useful to those dealing in the markets in question or to those concerned with their organization or regulation. The importance in the economy of these commodities justifies the further hope that the study will enhance our understanding of economic fluctuations in general. A great deal of statistical material is available which relates to the markets in question, especially for recent periods.

The theory on which this study is based is mainly concerned with the interrelations between the prices in the cash and futures markets. Its cornerstone is the so-called theory of "normal backwardation," first enunciated by the late Lord Keynes in his Treatise on Money. According to this theory, futures prices are normally below the spot price expected to rule at the time of delivery. This doctrine has been reformulated so as to free it from certain theoretical objections and to make it applicable to the conditions in the major American commodity markets. It has been tested empirically in various ways, particularly by studying the complex of relations between the level of stocks and the difference between cash prices and futures prices. (This difference is known as the "basis.") In this connection the influence of the government's price support programs has also been analyzed and found to be considerable.

Of special importance among these relations is that between the volume of hedging and the basis. An empirical example is given in Figure 1, which shows, for each month during the period October, 1946–September, 1952, the average value of the basis index for corn (a measure of the basis with adjustment for storage cost and interest) and the volume of "short" hedging by large corn traders. "Short" hedging is the more common form of hedging in which the trader owns actual merchandise and is short in futures. It will be seen that as a rule a large volume of short hedging is accompanied by a low value of the basis index. The opposite has been found to be true for "long" hedging, in which the trader has sold merchandise forward and holds futures as a hedge.

The question of who gains and who loses in the futures market has always intrigued observers. One implication of the theory of normal backwardation is that in the long run the hedgers lose money on the futures side of their transactions. This "loss" can more properly be looked upon as a payment to the speculators in return for the risk-bearing service they provide. To test this implication, the gains and losses of three groups of futures traders — the large hedgers, the large speculators, and the small traders have been estimated. It was found that on the whole the large hedgers do lose in the futures market, though of course these losses have to be made up in the cash market. The large futures speculators in the grain markets appear to gain not only from the hedgers but also from the small traders. The latter are for the most part outside speculators, not professionally associated with the market. This lack of success has also been noted by previous investigators; in the cotton market, however, it does not seem to be true that small traders lose on the average.

The second empirical study concerns the market in livestock and livestock products. It was conducted by Clifford Hildreth, assisted by Frank Jarrett, as a co-operative project with the Agricultural Economics Research Unit of the University of Chicago, and will appear shortly as Cowles Commission Monograph No. 15.

Livestock and livestock products — cattle, calves, hogs, chickens, turkeys, sheep, milk, and eggs — account for over half of the cash receipts from farming in the United States. In the form in which these products reach the consumer, they absorb over half of the consumer's expenditure for food. Fluctuations in prices and quantities of these commodities therefore have important effects on farm income and on the purchasing power of the consumer's dollar.

In throwing together several products into one category for purposes of analysis, there is always some loss of precision, in that one can no longer use one's knowledge with respect to demand or supply conditions for these individual commodities. Since the commodities in question are fairly good substitutes for each other both in production and in consumption, however, their prices retain fairly close relationships. For this reason the aggregation of these products is not so harmful as it otherwise might be, and it does bring the research task down to a level manageable within one project. The study may then be looked upon as a forerunner to more detailed analyses of individual types of livestock.

An economic model was drawn up containing the following relationships:

1. A relation explaining annual production of livestock (and products) on the basis of the quantities of feed grains, protein feeds, and roughage fed; inventory of livestock at beginning of the year; and a trend representing slow but sustained improvement in feeding efficiency
2. A relation explaining demand for feed grain on the basis of the prices of feed grain, of protein feeds, and of livestock (and products); the livestock inventory; and the amount of roughage fed
3. A relation explaining demand for protein feed from the same set of variables
4. A relation explaining supply of livestock (and products), that is, the quantity sold from farms, on the basis of the same three prices; the livestock inventory; and the price of farm labor
5. A relation explaining the demand for livestock (and products), measured by the same quantity as the supply, on the basis of the price of livestock (and products); consumers' income; population; and a cost-of-living index (excluding livestock and product prices)

To estimate these relationships from annual data for the period 1920–1949, a statistical model was adopted representing these relationships by a particular form of algebraic equations ("linear in the logarithms"), with allowance for the random elements in economic behavior. Various methods of estimation were tried out, including the so-called "limited information" method developed by the Cowles Commission. The results were compared, and statistical and other evidence bearing on the assumptions underlying alternative methods was examined.

An interesting finding is that, in the fourth supply relation above, a rise in the price of livestock and products diminishes current marketings. This possibility has been the subject of much conjecture and is not hard to explain, since higher current prices are generally associated with more favorable expectations about future prices. For most livestock, raising planned future production requires withholding more of the animals currently on hand for breeding and sending fewer to market. Thus current and future marketings compete in two ways. What is sold now cannot be sold later, and what is sold now cannot be used to produce more to be sold later. While these considerations do not prove that current marketings decrease with an increase in current price, they make it seem entirely possible. Assumptions of such a short-run negative response have played an important part in theories of instability in livestock markets, particularly for cattle, and it is interesting that they tend to be borne out by these results.

Application of the relationships to data for the year 1950 gave satisfactory "predictions" from all relationships except the fifth, the demand for livestock products. Increased competition between vegetable oils and animal fats was pinpointed as a likely contributing factor in the failure of the latter relation.

Ever since Adam Smith, economic theorists have been concerned with appraising the effectiveness of competitive markets as instruments for utilizing available resources for production and for distributing the product. It should not be thought that this is a closed subject. Recent advances in the introduction of new mathematical tools in economics have introduced greater clarity and rigor into this analysis. The Cowles Commission's work in this area during the period of this report was done largely by Gerard Debreu, partly in co-operation with Kenneth J. Arrow (related work by Hurwicz is described in Section 4 of this report).

The task of any productive and distributive system can be seen as follows. Let a typical consumer be presented with two or more alternative consumption programs, that is, two or more alternative statements specifying for each year in the present and in the future the flow of consumers' goods that will come to him in that year. Let it then be assumed that our consumer has a definite notion as to which consumption program would give him greater satisfaction. Let us further assume that each producer has a good knowledge of the alternative production programs, covering the present and the future, that are within the range of technical possibilities. (For the moment, let us rule out uncertainty in either future consumers' preferences or technological development.) There are, of course, limits to the amount of land that can be cultivated, to the rate at which minerals can be located and extracted, and to the amount of labor that can be rendered by a population increasing at a given rate. Finally, there is only a given amount of man-made equipment, buildings, and inventories available at the beginning of the period considered. The task of an economic system now is to utilize the available resources and the technological possibilities in such a way that in some sense maximum satisfaction is obtained there from by consumers over an indefinite period.

However, this statement needs clarification. It is always possible to give some consumers a higher satisfaction level by increasing their consumption at the cost of that of others. In order to evaluate an economic system fully, therefore, we would need some criterion for the fairness or appropriateness of the distribution of income that it brings about. The search for such a criterion leads economists into the difficult realm of ethical philosophy. No generally accepted criterion has so far been developed. Owing to the ingenuity of the Italian economist Pareto, however, a second line of defense has been established in the problem of appraising an economic system. It is possible to formulate a criterion that concentrates on productive and distributive efficiency only, while skirting the problem of distributive fairness. Following Pareto, we say that an economic system utilizes its resources efficiently if it affords each consumer a level of satisfaction that could not possibly be raised, within the given limits of technology and resource availability, except at the cost of lowering someone else's satisfaction. In such a case, there is at least no slack anywhere in converting resources into satisfaction. Of course, one can visualize many different income distributions, with each of which resources are utilized efficiently.

In the competitive, private-enterprise economy, no economic agent is much concerned with protecting the consumption or income levels of other agents. Each commodity has a price in each period; each consumer seeks to maximize his satisfaction level by proper allocation of his expenditure in each period to the various commodities available at those prices; and each producer seeks to use his knowledge to maximize his profit, that is, the excess of the receipts from sale of products at those prices over the cost of the various inputs procured at those prices. Finally, prices are such as to equate supply and demand. Now the remarkable fact is that, if competition is perfect, that is, if everyone takes prices as given without attempting to influence them by the amounts offered or withheld from sale, then this competitive economy does use resources efficiently. This proposition, foreseen in Adam Smith's reference to the "invisible hand," and first explicitly formulated by Pareto, is fully and rigorously proved in the work here reported on.

Pareto introduced also another and perhaps even more interesting idea, which leads to a second proposition of which the proof has been completed. Suppose there is an economy which does not use the idea of pricing all commodities and services. Instead, for instance, all production and distribution is carried out on orders from an all-powerful government. Suppose further that, by some unexplained miracle of planning, this economy does manage to use its resources efficiently, in the same sense indicated above. Then, so says our second proposition, whether the people involved know it or not, there is a system of prices such that, if each consumer maximized his satisfaction at those prices, and if each producer maximized his profit at those prices, the economy would be in precisely the same state. Somewhat over-simplified: the perfectly competitive market utilizes resources as well, insofar as productive and distributive efficiency is concerned, as the theoretical maximum of perfect planning can do. Since, in fact (and quite apart from any consideration of individual freedom), "perfect" planning without use of prices would require an accumulation of technical information and computation equipment that far surpasses the capacity even of modern electronic devices, competitive markets and the price system must be looked upon as invaluable instruments for the efficient utilization of resources, permitting allocative decisions to be decentralized and spread out over a great many enterprises.

The term "proof" has been used above. It should be said at once that any "proof" refers to a mathematical model of some reality. The reality itself is always more complicated. For instance, the model discussed disregards uncertainty, about which a few remarks below. The model also disregards the fact that some resources or products come in indivisible units, a fact which is at the root of the greater productivity of mass production. It also disregards production processes which involve harm (such as make pollution) or benefit (such as the combating of insect pests) to others than those controlling the production decisions. But, if we grant all the simplifying assumptions made, the proofs are rigorous from there on. In fact, the hardest part of the mathematical analysis is not immediately related to the propositions mentioned. It is concerned with proving the logical compatibility of the assumptions that make up the model of a competitive economy. It needs to be proved that there exists an equilibrium. That is, given the preferences of consumers, the technology of production and the resource availabilities, there is always a set of prices at which, if each consumer maximizes his satisfaction, and each producer his profits, equilibrium of supply and demand results in each market. By building upon work of Abraham Wald and of John von Neumann, such a proof has been supplied for the model of a competitive economy described above.

Extensions of these studies to take into account uncertainty concerning future technology, resource availability, and preferences have also been explored. It may be expected that some results carry over from the case of certainty to that of uncertainty. However, any efficiency properties of competitive markets in the case of uncertainty are likely to depend on whether the anticipations concerning the future held by the various decision-makers are the best that can be made and are mutually consistent.

2. Study of Transportation Systems

In earlier reports we have described studies of a model of transportation which was sufficiently simple to have some relevance for all transportation systems that depend on moving vehicles, ships, or other carriers. During the period of this report more detailed studies were made which related more specifically to two particular transportation systems: highway traffic and railroad transportation. Beckmann, Koopmans, McGuire, Winsten, and, during parts of the period, Nerlove and Goldman formed the team working on these problems, under Koopmans' direction.

Highway transportation has many features in common with markets. Traffic flows result from the interaction of many individual decisions with. regard to whether or not to travel by road, which route to take, and at which speed. Apart from toll roads, however, no price is charged as a condition for the use of any particular road at any particular time. The "price" of road use is incurred indirectly, in the form of travel time, fuel cost, vehicle depreciation, etc. Interaction between decisions arises particularly in the case of congested roads, where traffic is dense enough to cause the cost of travel to any one road user to be dependent on the number of other simultaneous road users.

As explained already, the kind of physical interaction where one man's decision places obstacles in another man's path was not considered in the general theory of competitive markets described above. Situations of this kind, however, have received considerable attention in economic literature. They are usually referred to as situations in which the social cost (the total cost to all individuals) of a decision exceeds the private cost borne by the deciding individual. With regard to highway traffic, the consequence of this circumstance is to increase congestion on the most traveled roads beyond the point of most efficient utilization of the highway network. Because normally only one's own cost is considered in choosing between a short congested road and a longer uncongested one, more drivers choose the congested road than would correspond to efficient use of the roads. Theoretically, a better utilization of the road network would be obtained if there were a way to collect from each user of a congested road a tax or toll which would act in some degree as a deterrent to the use of that road. If these toll rates were properly determined, the drivers that remained on a road would be those to whom the value of the use of that road was at least as great as the total cost (other than tolls) borne by all road users on their account. To prevent the tolls from becoming a net deterrent to all road use, the revenue collected could be applied to benefit all road users, for instance, by lowering gasoline taxes. Our studies have led us to set up equations from which such "efficiency toll rates" can be determined, but they have not led to practical suggestions as to how such tolls could be collected. However, they also indicate an answer to the more limited problem of how to set the rates on roads that are already toll roads in such a way as to promote better utilization of the highway network. They also give us a start on the problem of deciding when it is worth while to build or improve a certain road, even when tolls cannot be collected.

Best utilization of the road network is not the only preoccupation of our study of highway transportation. Methods for predicting highway traffic resulting from changes in capacity are also developed and discussed. In addition, the mathematical theory of waiting lines is applied to provide background to the analysis of congestion in various situations, such as at intersections or on two-lane two-way roads.

Compared with the wide dispersal of decision-making in high-way traffic, a railroad company is a highly controlled entity, in theory subject to one policy determined by management and implemented all the way down the line of command. At the same time the technology of railroad operation is a good deal more complicated. To illustrate the difference, consider the fact that the engineer cannot in all circumstances with his own eyes ascertain that the track ahead is free at least for a distance within which his train can be stopped. This fact alone already necessitates a great deal of organization and central control to insure correct handling of the information on which safety depends. With the predominance of central control goes a slight shift in the purposes and uses of scientific analysis. In regard to highway traffic, the emphasis is on predicting what many independent drivers will do. In railroad transportation, greater emphasis falls on the type of analysis that can help management to make decisions resulting in greater efficiency of operation. Thus we enter an area of research, variously called "management science," "operations research," or "industrial engineering," concerned with the study of operational efficiency. The last few years have seen a considerable development in this field.

The Commission's exploratory studies of railroad operation were not performed for a railroad company or other railroad industry group.(+) Their original motivation was to contribute to the development of concepts and models that would help in defining and assessing the capacity of a railroad network. It soon became clear that this problem presupposes analysis of the main aspects of railroad operation, such as classification, or switching, policies at different freight yards, congestion delays prior to classification, "accumulation delays" of cars waiting after classification until enough cars are available to form a train, scheduling of trains on connecting lines, etc. Besides an exploratory discussion of the main areas of railroad operation, this work has led to a few models for the discussion of a few specific operational problems, such as the assignment of classification work to successive yards down the line and the scheduling of trains on a single line. A report on the studies of highway and railroad transportation is being prepared for publication.

3. Theory of Organization

The comparison of highway and railroad transportation has shown how differences in technology between two systems serving quite similar purposes can lead to entirely different forms of organization. The need for a theory of organization is thereby suggested. How does the most suitable form of organization depend on the task to be performed and on the technological characteristics of the means to be used? How does interaction between decision-makers within an organization differ from interaction through a market or from interaction through sharing the use of facilities such as roads?

The theory of organization has comprised, since 1952, a major area of research at the Cowles Commission. This research is carried out by Jacob Marschak, Roy Radner, and recently Richard F. Muth and Donald Bratton, with Marschak as the project leader. The approach that is followed can perhaps best be illustrated by a model which is simple and admittedly unrealistic, yet which may be useful. Bodies falling in a vacuum, or bridge-frames conceived as webs of geometric lines, are also unrealistic models of real-world phenomena. Yet they proved essential for practical engineering by clarifying certain basic principles that could be applied to complicated reality. We hope to clarify in a similar way the essential principles for efficient organization forms, in private business as well as in the civilian or military agencies of the government.

Imagine a company engaged in shipping a product between two regions (call them North and South), buying where it is cheap and selling where it is dear. The company may hire market specialists (observers) who can, we shall suppose, at the time purchase or sales orders are given, predict unfailingly the price in their respective markets; and it can hire personnel for shipping and trading operations. Suppose the firm's shipping facilities are limited, at each terminal point, to one trainload per week and let the cost of transporting one trainload be the unit of value in which all other costs or prices will be expressed. Suppose the firm knows (without the services of a market specialist) that the product price per trainload in the North is equally likely to be 6 or 10 or 14 units of value and that in the South the price is equally likely to be 7 or 8 or 9 units, independently of what the northern price is. What kind of organization will, in the long run, profit the company most? For example: Shall there be a specialist for each market, for one market, or for none? Shall all market information flow to one decision center or to two local branches? Shall these decide independently whether or not to ship (thus occasionally shipping in two opposite directions at the same time), or shall time and money be invested in consultations (periodic or occasional) between the branches? Or shall one branch take orders from the other?

The forms of organization can be grouped according to which decision is based on what information. In our example, the decision whether or not to ship from North to South can be based on either (1) the knowledge of both prices, or (2) the knowledge of the North price, or (3) the knowledge of the South price, or (4) the knowledge of price probabilities only. The decision whether or not to ship from the South to the North can be similarly based on any of the four different kinds of knowledge listed. Hence, in our example, all organization forms can be grouped into 4 x 4 =16 classes. In particular, a centralized organization would belong to the class where each decision is based on the knowledge of both prices. Such an organization would reap, in the long-run average, a profit that could not be lower than that of any decentralized organization, if the time of the central decision-maker were costless and the communication and processing of all the data were also free of cost. If, for purposes of a step-wise analysis, the reader postpones the consideration of these "organization costs," he can compute, for each organization class, the "maximum expected gross profit" that will be obtained in the long-run average if each decision-maker always uses his information according to a "good" (i.e., on the average, most profitable) decision rule. This is used in the table on the following page, where selected organization charts are grouped into only 4 classes. The remaining 16 – 4 = 12 classes are easily seen to be ineligible as they require more information, and hence presumably higher organization costs, without yielding a higher expected gross profit than the 4 listed classes. The reason for this lies in the relatively small range of variation of the southern price.

Note, in the lower right corner, a "routine" organization: one shipping agent performs the same operation every week, sometimes incurring a loss, but earning a modest gross profit on the average, against which are presumably debited only modest organization costs. The less routine there is and the more fresh information is used, the larger the expected gross profit, but also the more complicated the organization and the higher its cost. The latter depends on the number and kind of positions (essentially, salaries of executives); on the number and kind of communication links (telephone, traveling); and on the cost of delays in transmission through successive links, or, more important, through congestion at the nodes (i.e., on the executives' desks). To replace "routine" organization by one in which, for example, the decision to ship in either direction is based on the northern price — the case in the upper left corner would only pay if the links, delays, and positions (other than that of the southern agent) involved in at least one organization chart of that class would cost less than 2-1/3 – 1 = 1-1/3 units. By thus taking organization costs into account and, in effect, comparing net rather than gross expected profits, the more efficient organization forms can be chosen.

In several respects the studies have gone beyond the features of the primitive example just given. For instance, the external events (such as, in our example, the two local prices) may include any number of variables, and these may be correlated instead of independent. The reader will guess the consequence: if two events are strongly correlated, knowledge of one of the two may be sufficient. Thought has also been given to the realistic case in which the probabilities of external events are not known beforehand: the team of executives has to make decisions while still learning about the relevant probabilities by accumulating and interchanging experience, thus gradually approaching the good decision rules and the maximum profit.

Another aspect is the assumption about how profit is determined by both the decisions and the external events. In the above example of purely speculative trading, each decision (to ship northward, southward) adds an independent contribution to the total profit. In reality, a company's profit can be seldom represented as a sum of the contributions of individual decision-makers, for example, of the production manager and the sales manager, or of the raw-materials purchaser and the personnel manager. Economists have been aware of this and have therefore usually represented a firm's output as some function (but usually not just a weighted sum) of its several inputs. In a sense, our objective is to extend the classical theory of the firm from the case of a single entrepreneur acting in full knowledge of all data to the case of a team of executives interchanging parts of their information and hence each acting under partial uncertainty.

In our example the decision-maker was ignorant of some variables (the price in a certain locality) while completely informed of others. In reality, all information the executive receives is incomplete. Before reaching him, the report is processed, reduced from details to essentials. Sometimes an executive transforms his information into a (presumably "good") decision; this he communicates to a "subordinate" executive in the form of an "order," or an "assignment"; the subordinate, in turn, with the help of additional, more special or local, information of his own, has then to choose a "good" decision on how to accomplish the assignment. Reports and orders are various forms of processed information and have obvious advantages over unprocessed masses of data. The concept of organization form (which, as we have seen, determines the maximum expected gross profit as well as the organization cost) must therefore include not only the network of executive positions and communications but also the way in which information is processed, "coded."

Perhaps one is permitted to assume that, in the course of decades of competitive struggle, business firms, or at least the more successful business firms, have already evolved reasonably efficient communication networks and codes, appropriate to the particular branch of industry or commerce and to other individual circumstances. If so, theory can be checked against facts.

Accordingly, the Cowles Commission's staff is having discussions with management consultant firms to draw on their experience; in addition, one large manufacturing company has kindly offered a desk to a Commission staff member, in exchange for some work on their problems of production scheduling and inventory control. Our staff member will seek to answer these questions: What information does each executive receive and give during a typical day, from whom and to whom? What decisions does he make? What time elapses between various communications and decisions? We hope this will help to put some flesh on the bare bones of the theory.

So far, we have spoken of teams of decision-makers. But the analysis of teams presupposes a study of a simpler and more basic problem: decisions of single persons. Such study does, in fact, form another part of the project on "Decision-making under Uncertainty." It has been pursued by Chernoff, Debreu, Herstein (in co-operation with John Milnor of Princeton), Marschak, Radner, and Tornqvist.

It may be questioned whether maximum expected profit (in dollars) is properly considered a decision-maker's goal. A more general criterion has been analyzed, that of "maximum expected utility," as is possibly exemplified by the strategist's concepts of "average military worth" and "calculated risk." With the help of the concept of maximum expected utility, one can again define "good rules" for the decision-maker. Such rules have been extended to the realistic case when one does not know in advance the probabilities of relevant events but can gradually improve his decisions helped by the inflow of successive observations. Measurable concepts such as value of information and value of precision have been developed.

Furthermore, members of the research group are interested in how ideally "good" decision-making, based on consistent goals and quiet deliberation, compares with habits of actual decisions by men in our society — decisions that are often made under pressure of time and are sometimes contradictory — and how the practice of good decision-making could be developed by training. For such studies there has been occasional co-operation with psychologists. One effect of this co-operation may be mentioned here: in line with the results of psychological experiments, we relax the assumption of consistent decision-making and assume that a man's choices are subject to variations; we continue to assume, however, that these choices obey definite probability laws and can therefore be studied statistically.

While the theory of organization and decision-making is still in its beginnings, it is believed that continued work in this field will ultimately find practical use in business as well as in governmental administration.

4. Other Substantive Research Topics

The research of the Cowles Commission in the period of this report has included, besides the major projects described, a number of additional investigations on a variety of problems. Some of these continue with ideas developed in earlier periods, and some of these are likely to be extended beyond the present period. We shall now indicate a few of these by brief comments, which of necessity will involve here and there some technical terminology. We shall in some cases refer to studies made by research consultants of the Commission in other universities, sometimes as a part of the research program of another institution. We shall first comment on substantive research topics not yet covered, and thereafter on studies aimed at the development of statistical or mathematical tools of analysis. In neither category will complete coverage be attempted. The list of titles of papers or addresses following this report may serve as supplementary information.

In our preceding report we mentioned a study by Klein, published during the present report period, relating to the input–output analysis of Professor Leontief of Harvard University. This analysis is designed to represent the effects upon output in each industry of given changes in final demand. Arrow has considered a problem suggested by an input–output analysis of the Italian economy by H.B. Chenery. The problem is to meet a given domestic demand from domestic production, limited by given industrial capacities, and from imports, in such a way as to minimize foreign exchange requirements. Surprisingly, he finds that, within wide limits, the best balance of imports and domestic production is independent of import prices.

Harberger made a study of the elasticity of United States demand for imports, using short-cut estimation methods, which are described in Section 5. His estimates of these elasticities are higher than those reached in previous studies, to an extent significant for policy conclusions. The discrepancy appears to be the result of the choice of a method of estimation that recognizes the existence of simultaneous economic relations between the variables involved.

Related to the theoretical analysis of competitive markets, described above, is work by Hurwicz on the possibilities of attaining efficiency in the use of resources by decentralized decisions guided by a price system. This study is an attempt to go beyond the assumption of perfect divisibility of resources and thus, for instance, to recognize the superior productivity of large units or complexes of machinery.

Another attempt to deal with allocation problems of indivisible resources is a study of the so-called "assignment problem," to which Beckmann, Koopmans, Motzkin, and Tornqvist have contributed and in which two visits from Professor Harold Kuhn of Bryn Mawr College have been a stimulating factor. The problem concerns assignment of personnel to jobs, or plants to locations, or any other x's to y's, in such a way as to maximize the economic benefit from the assignment. A price system which sustains an optimal assignment has been found for the case in which the benefit from each x–y-combination is independent of how other x's and y's are matched. Since this condition is not generally met in the assignment of plants to locations, the efficiency of locational choices guided by the price system is as yet in question.

Hildreth and Houthakker both made studies in which the concepts of activity analysis were used or extended to analyze alternative production possibilities in one industry. The former discussed the choice of a fertilizer treatment in cotton-growing in response to market prices of the product and of factors of production. The latter laid a connection between the variability of input–output ratios as between establishments in an industry and the aggregate production function of that industry.

5. Statistical and Mathematical Tools

During the period of this report Cowles Commission Monograph 14, Studies in Econometric Method, was published. It summarizes the main results of the Commission's work in the statistical methodology of the measurement of economic relations, published earlier in more technical presentation. In addition, it contains new work on methodological problems, described in our preceding report.

Monograph 14 is addressed in the main to statisticians interested in econometrics. Another book, in preparation by Carl Christ, addresses the general economist in expository language. It deals with economical model construction in general as well as with the associated statistical estimation problems.

Studies in statistical methods for econometrics during the period of this report have been in the nature of explorations and new departures. Hildreth has dealt with the problem of estimating a production function when it is known that successive equal increases in input produce successively smaller (or equal) increases in output. Harberger has demonstrated the feasibility of computationally very simple estimating procedures. On the basis of explicitly stated assumptions about the possible extent of year-to-year shifts in demand or supply functions, he arrives at surprisingly narrow limits on the elasticities of demand or supply in question. Houthakker has discussed how the specification of the variables that one can afford to include in estimating a behavior relationship depends on the number and kind of observations available.

Winsten and Prais studied the estimation of trends in the case where the "deviations from trend" in a time series are serially correlated. Both Radner and T.W. Anderson, independently, studied the possibility of basing the analysis of time series and relations between time series on a continuous time variable rather than a discrete one.

With regard to mathematical tools used in economics, horizons are expanding. An expository article by Herstein, addressed to the applied mathematician, gives a variety of examples of mathematical techniques which economists have found useful. Debreu has given a summary of separation theorems of convex sets, which are crucial to the understanding of the role of prices as guides to the allocation of resources. Beckmann has discussed inequalities that indicate how simultaneous changes in technology, in resource availability, and in valuations of end-products affect rates of output and prices in various industries. These extend earlier results by Samuelson known as the "le Chatelier principle of linear programming." Davis has continued his systematic studies of nonlinear operators and differential equations.

6. Concluding Remarks

At the beginning of this report we have emphasized the fascination which economic research holds for those engaged in it. We do not know whether the description of specific pieces of research has conveyed a feeling for the challenge of the problems concerned. However, we do not wish to leave the impression that intellectual delight in the subject matter is regarded as the ultimate motivation for engaging in economic research. The relation of ends and means is just the other way around. There is ample precedent, in the natural and social sciences alike, for the belief that a free reign to intellectual curiosity may open up the pathways to useful and applicable knowledge more readily than a carefully charted plan of approach. In the selection of problems for investigation and of methods of study, the criterion of intellectual stimulation to open minds takes at least an equal place with any other relevant criteria.

SELECTED LIST OF STAFF PUBLICATIONS, PAPERS, AND ADDRESSES

1. Studies of Markets

• Arrow, Kenneth J. "Le Role des valeurs boursieres pour la repartition la meilleure des risques," Colloque international d'econometrie, 1952, pp. 41–47. Paris: Centre National de la Recherche Scientifique, 1953. (Reprinted as CCNS, No. 77.)
• Arrow, Kenneth J., and Debreu, Gerard. "Existence of an Equilibrium for a Competitive Economy," Econometrica, Vol. 22, July, 1954, pp. 265–290. (Reprinted as CCNS, No. 87.)
• Beckmann, Martin J. "A Note on the Theory of Forward Markets." CCDP, Economics 2066.
• Debreu, Gerard. "A Social Equilibrium Existence Theorem," Proceedings of the National Academy of Sciences of the U.S.A., Vol. 38, October, 1952, pp. 886–893. (Reprinted as CCNS, No. 64.)
• ——. "Une economie de l'incertain," presented December 9, 1953, at the Ecole Nationale Superieure des Mines, Paris, France.
• ——. "Valuation Equilibrium and Pareto Optimum," Proceedings of the National Academy of Sciences of the U.S.A., Vol. 40, July, 1954, pp. 588–592. (Reprinted as CCNS, No. 84.)
• Hildreth, Clifford, And Jarrett, Frank. A Statistical Study of Livestock Production and Marketing. Cowles Commission Monograph No. 15. New York: John Wiley & Sons (in press).
• Houthakker, H.S. "Stocks and Spreads in Future Markets," presented September 2, 1952, before the Econometric Society, East Lansing, Michigan.
• ——. "Commodity Futures" (with L.G. Telser), presented December 28, 1953, before the Econometric Society and the American Farm Economic Association, Washington, DC.
• ——. Commodity Futures: A Study in the Economics of Uncertainty (with the assistance of Lester G. Telser). (To be published as a Cowles Commission Monograph.)
• Telser, Lester G. "Safety First, Short Hedging, and the Commodity Credit Corporation." CCDP, Economics 2104.
• ——. "Application of Safety First to Hedging." CCDP, Economics 2107.

2. Study of Transportation Systems

• Beckmann, Martin J. "A Continuous Model of Transportation," Econometrica, Vol. 20, October, 1952, pp. 643–660. (Reprinted as CCNS, No. 66.)
• ——. "Efficient Transportation in Networks," presented September 3, 1953, before the Econometric Society, Kingston, Ontario, Canada.
• Beckmann, Martin J., McGuire, C. Bartlett, and Winsten, Christopher B. "Studies in the Economics of Transportation," with an introduction by Tjalling C. Koopmans (forthcoming).
• Fox, Kirk. "Economical Routing of Empty Railroad Freight Cars." CCDP, Economics 2047.
• McGuire, C. Bartlett. "The Allocation of Switching Work in a System of Classification Yards" (with M. Beckmann, T.C. Koopmans, and C.B. Winsten), Proceedings of the Railway Systems and Procedures Association, Winter Meeting, Chicago, Illinois, November 4, 5, 6, 1953, published by the Railway Systems and Procedures Association, December, 1953.
• ——. "A Theoretical Study of Traffic Congestion" (with C.B. Winsten), presented May 16, 1953, before a meeting of the Operations Research Society of America, Cleveland, Ohio. Abstract in Journal of the Operations Research Society of America, Vol. 1, May, 1953, pp. 149–150.
• Nerlove, Marc. "Optimal Routings of Empty Boxcars, 1940–1950," CCDP, Economics 2086.

3. Theory of Organization

• Beckmann, Martin J. "On Marschak's Model of an Arbitrage Firm" (with Daniel Waterman), presented September 4, 1952, before the Econometric Society, East Lansing, Michigan. Also CCDP, Economics 2058.
• ——. "On the Optimal Path of Expansion of a Firm." CCDP, Economics 2101.
• Bratton, Donald. "The Arbitrage Problem Where Each Partner Can Find Himself in One of Two Situations and Can Perform One of Two Actions." CCDP, Economics 2075.
• ——. "Efficient Communication Networks." CCDP, Economics 2102.
• Chernoff, Herman. "Rational Selection of Decision Functions," Econometrica, Vol. 22, October, 1954, pp. 422–443. (To be reprinted as CCNS, No. 91.)
• Debreu, Gerard. "Representation of a Preference Ordering by a Numerical Function," chap. XI in Decision Processes (Thrall, Davis, and Coombs, eds.), pp. 159–165. New York: John Wiley & Sons, 1954.
• Faxen, Karl O. "Incentive Functions and the Arbitrage Firm." CCDP, Economics 2059.
• Herstein, I.N., and Milnor, John. "An Axiomatic Approach to Measurable Utility," Econometrica, Vol. 21, April, 1953, pp. 291–299. (Reprinted as CCNS, No. 65.)
• Kiefer, J., and Grey, S. "On the Arbitrage Problem." CCDP, Economics 2068.
• Koopmans, Tjalling C. "La Notion d'utilité clans le cas de décisions concernant le bientitre futur," Cahiers du séminaire d'économétrie (RENE Roy, ed.), No. 2, pp. 7–10. Paris: Centre National de la Recherche Scientifique, 1953.
• Marschak, Jacob. "Équipes et organisations en régime d'incertitude," Colloque international d'économétrie, 1952, pp. 201–211. Paris: Centre National de la Recherche Scientifique, 1953.
• ——. "Mathematical Models and Experiments on Decision-making," presented February 26, 1954, before a meeting sponsored by the Armed Forces Epidemiological Board, Office of the Surgeon General, U.S. Army; and Dunlap and Associates. Also CCDP, Economics 2097.
• ——. "Probability in the Social Sciences," in Mathematical Thinking in the Social Sciences (Paul F. Lazarsfeld, ed.), pp. 166–215. Glencoe, Ill.: Free Press, 1954. (Reprinted as CCNS, No. 82.)
• ——. "Towards an Economic Theory of Organization and Information," chap. XIV in Decision Processes (Thrall, Davis, and Coombs, eds.) pp. 187–220. New York: John Wiley & Sons, 1954.
• ——. "Toward a Preference Scale for Decision-making," in Readings in Game Theory and Political Behavior (M. Shubik, ed.), pp. 22–32. Doubleday Short Studies in Political Science. Garden City, NY: Doubleday & Co., Inc., 1954.
• Marschak, Jacob, and Mickey, R.M. "Optimal Weapon Systems," Naval Research Logistics Quarterly, Vol. 1, June, 1954, pp. 116–140.
• Marschak, Jacob, and Radner, Roy. "Structural and Operational Communication Problems in Teams, I and II." CCDP, Economics 2076, 2077.
• ——. "The Firm as a Team," presented December 29, 1953, before the Econometric Society and the Institute of Mathematical Statistics, Washington, D.C. Abstract in Econometrics, Vol. 22, October, 1954, p. 523.
• ——. "Note on Some Proposed Decision Criteria," chap. v in Decision Processes (Thrall, Davis, And Coombs, eds.), pp. 61–68. New York: John Wiley & Sons, 1954.
• Radner, Roy. "On Optimal Communication Rules for Certain Types of Teams." CCDP, Economics 2064.
• Radner, Roy, and Tritter, A. "Communication in Networks." CCDP, Economics 2098.
• Simon, Herbert A. "A Comparison of Organisation Theories," Review of Economic Studies, Vol. 20, No. 51, 1952–1953, pp. 40–48. (Reprinted as CCNS, No. 47.)
• ——. "A Behavioral Model of Rational Choice," June 23, 1953 (unpublished hectographed paper).
• Tornqvist, Leo. "Some Game Theoretic Points of View on Scientific Research." CCDP, Economics 2056. Also presented December 27, 1952, before the Econometric Society, Chicago, Illinois.
• ——. "A Theory for the Value of a Message and Systems for Codifying Possible Messages." CCDP, Economics 2060.
• ——. "The Problem of Finding Optimal Decisions." CCDP, Economics 2062.
• ——. "A General and Realistic Approach to the Decision Problem," presented September 4, 1952, before the Econometric Society, East Lansing, Michigan.
• ——. "The Best Can Be an Enemy of the Good." CCDP, Economics 2074.
• ——. "Introduction to a Study of Decision-making." CCDP, Economics 2080.
• ——. "The Problem of Constructing Indicators of Goodness." CCDP, Economics 2081.
• ——. "Decision-making in a Nonneutral World." CCDP, Economics 2084.
• ——. "A Model for Stochastic Decision-making." CCDP, Economics 2100.
• Tornqvist, Leo, And Marschak, Jacob. "Conditions under Which Communication Is Superfluous." CCDP, Economics 2078.

4. Other Substantive Research Topics

• Allen, Stephen G. "Inventory Fluctuations in Flaxseed and Linseed Oil, 1926–1939," Econometrica, Vol. 22, July, 1954, pp. 310–327. (Reprinted as: CCNS, No. 86.)
• Arrow, Kenneth J. "Le Principe de rationalité clans les choix collectifs," Economic appliqueé, Vol. 5, October–December, 1952, pp. 469–484.
• ——. "The Determination of Many-Commodity Preference Scales by Two-Commodity Comparisons," Metroeconomica, Vol. 4, December, 1952, pp. 105–115. (Reprinted as CCNS, No. 62.)
• ——. "Import Substitution in Leontief Models," presented December 27, 1953, before the Econometric Society, Washington, DC. Also Econometrica, Vol. 22, October, 1954, pp. 481–492. (To be reprinted as CCNS, No. 90.)
• Beckmann, Martin J. "Bemerkungen zum Verkehrsgesetz von Lardner," Weltwirtschaftliches Archiv, Vol. 69, No. 2, 1952, pp. 199–215.
• ——. "The Partial Equilibrium of a Continuous Space Market," Weltwirtschaftliches Archiv, Vol. 71, No. 1, 1953, pp. 73–89.
• ——. "Bemerkungen zu Bombachs Modell des Wirtschaftlichen Wachstums," Weltwirtschaftliches Archiv, Vol. 72, No. 1, 1954, pp. 102–105.
• Beckmann, Martin J., and Koopmans, T.C. "A Note on the Optimal Assignment Problem." CCDP, Economics 2053.
• ——. "On Some Assignment Problems," presented May 16, 1953, before a meeting of the Operations Research Society of America, Cleveland, Ohio. Abstract in Journal of the Operations Research Society of America, Vol. 1, No. 3, 1953, p. 149.
• Beckmann, Martin J., and Marschak, Thomas. "On the Theory of Location in the Short Run." CCDP, Economics 2099. Also reprinted in the Technical Report prepared for the Office of Naval Research (NR–047–004), Stanford University, March 26, 1954.
• Christ, Carl. "A Review of Input–Output Analysis," presented October 17, 1952, before the Conference on Research in Income and Wealth, New York. (Forthcoming in Studies in Income and Wealth, National Bureau of Economic Research, Vol. 18.)
• Debreu, Gerard. "A Classical Tax-Subsidy Problem," Econometrica, Vol. 22, January, 1954, pp. 14–22. (Reprinted as CCNS, No. 80.)
• ——. "Numerical Representations of Technological Change," Metroeconomica, Vol. 6, August, 1954, pp. 45–54. (To be reprinted as CCNS, No. 83.)
• Haavelmo, Trygve. A Study in the Theory of Economic Evolution. A monograph published by North-Holland Publishing Company, Amsterdam, 1954. 114 pp.
• Harberger, Arnold C. "A Structural Approach to the Problem of Import Demand," Papers and Proceedings of the American Economic Review, Vol. 43, May, 1953, pp. 148–159. (Reprinted as CCNS, No. 73.)
• Hildreth, Clifford C. "Ethical Criteria for Group Choice: A Preliminary Formulation." CCDP, Economics 2063.
• ——. "A Suggested Approach to Problems of Group Choice," presented December 29, 1952, before the Econometric Society and Institute of Mathematical Statistics, Chicago, Illinois.
• ——. "Alternative Conditions for Social Orderings," Econometrica, Vol. 21, January, 1953, pp. 81–94. (Reprinted as CCNS, No. 68.)
• ——. "Economic Implications of Some Cotton Fertilizer Experiments," to be published in Econometrica, Vol. 23, January, 1955.
• Houthakker, H.S. "Aggregative Activity Analysis." CCDP, Economics 2085.
• ——. "Demand Analysis — Achievements and Prospects," presented September 2, 1953, before the Econometric Society, Kingston, Ontario, Canada.
• ——. "La Forme des courbes d'Engel," Cahiers du séminaire d'économétrie (Rene Roy, ed.), No. 2, pp. 59–66. Paris: Centre National de la Recherche Scientifique, 1953.
• ——. "The Pareto Distribution and the Cobb–Douglas Production Function in Activity Analysis," to be published in the Review of Economic Studies.
• Hurwicz, Leonid. "Decentralized Resource Allocation." CCDP, Economics 2070.
• ——. "What Has Happened to the Theory of Games," Papers and Proceedings of the American Economic Review, Vol. 43, May, 1953, pp. 399–405. (Reprinted as CCNS, No. 75.)
• ——. "Programming in General Spaces." CCDP, Economics 2109.
• Klein, Lawrence R. "On the Interpretation of Professor Leontief's System," Review of Economic Studies, Vol. 20, No. 52, 1952–1953, Pp. 131–136. (Reprinted as CCNS, No. 69.)
• Klein, Lawrence R., and Mooney, H.W. "Negro–White Savings Differentials and the Consumption Function Problem," Econometrica, Vol. 21, July, 1953, pp. 425–456.
• Koopmans, Tjalling C. "Activity Analysis and Its Applications," Papers and Proceedings of the American Economic Association, Vol. 43, May, 1953, pp. 406–415. (Reprinted as CCNS, No. 75.)
• ——. "Uses of Prices," Proceedings of the Conference on Operations Research in Production and Inventory Control. Cleveland: Case Institute of Technology, 1954. (Reprinted as a Cowles Commission Special Paper, No. 3.)
• Malinvaud, Edmond. "Capital Accumulation and Efficient Allocation of Resources," Econometrica, Vol. 21, April, 1953, pp. 233–268. (Reprinted as CCNS, No. 71.)
• Marschak, Jacob. "Monnaie et liquidité dans des modèles macroéconomiques et microéconomiques," Cahiers du séminaire d'économétrie (Rene Roy, ed.), No. 3. Paris: Centre National de la Recherche Scientifique, 1954 (forthcoming).
• Motzkin, Theodore S. "The Assignment Problem." CCDP, Mathematics 425. (To be published in the Proceedings of the Sixth Symposium for Applied Mathematics, American Mathematical Society.)
• Simon, Herbert A. "Notes on Two Approaches to the Production Rate Problem." CCDP, Economics 2057.
• Strotz, R.H., Mcanulty, J.C., and Naines, J.B., Jr. "Goodwin's Nonlinear Theory of the Business Cycle: An Electro-Analog Solution," Econometrica, Vol. 21, July, 1953, pp. 390–411. (Reprinted as CCNS, No. 74.)
• Tornqvist, Leo. "How To Find Optimal Solutions to Assignment Problems." CCDP, Mathematics 424.
• Yasui, Takuma. "Nonlinear Self-excited Oscillations and Business Cycles." CCDP, Economics 2065.

5. Statistical and Mathematical Tools

• Anderson, T.W. "On Estimation of Parameters in Latent Structure Analysis," Psychometrika, Vol. 19, March, 1954, pp. 1–10.
• Beckmann, Martin J. "A Lagrangian Multiplier Rule in Linear Activity Analysis and Some of Its Applications." CCDP, Economics 2054.
• ——. "The Generalized (Weak) Le Chatelier Principle in Linear Activity Analysis." CCDP, Economics 2092.
• ——. "On a Variational Problem in Nonstatic Linear Activity Analysis." CCDP, Economics 2095.
• Christ, Carl F. "What Kind of Data for Econometrics" (abstract), Econometrica, Vol. 21, April, 1953, pp. 338–339.
• ——. "Pitfalls in Econometrics," presented December 28, 1953, before the Econometric Society and the American Economic Association, Washington, DC.
• ——. An expository monograph on econometric models and methods, to be published as a Cowles Commission monograph.
• Debreu, Gerard. "The Continuity of Multivalued Functions in Economics. CCDP, Economics 2079.
• ——. "Separation Theorems for Convex Sets." CCDP, Mathematics 423.
• Debreu, Gerard, and Herstein, I.N. "Nonnegative Square Matrices," Econometrica, Vol. 21, October, 1953, pp. 597–607. (Reprinted as CCNS, No. 76.)
• Gurland, John. "An Example of Autocorrelated Disturbances in Linear Regression," Econometrica, Vol. 22, April, 1954, pp. 218–227. (Reprinted as CCNS, No. 85.)
• ——. "Distribution of Semidefinite and of Indefinite Quadratic Forms" (abstract), Annals of Mathematical Statistics, Vol. 24, March, 1953, p. 138.
• ——. "Distribution of Quadratic Forms and Ratios of Quadratic Forms," Annals of Mathematical Statistics, Vol. 24, September, 1953, pp. 416–427. (Reprinted as CCNS, No. 79.)
• Harberger, Arnold C. "On the Estimation of Economic Parameters." CCDP, Economics 2088.
• Herstein, I.N. "Comments on Solow's 'Structure of Linear Models'," Econometrica, Vol. 20, October, 1952, pp. 685–686.
• ——. "Some Mathematical Methods and Techniques in Economics," Quarterly of Applied Mathematics, Vol. 11, October, 1953, pp. 249–262. (Reprinted as CCNS, No. 78.)
• Hildreth, Clifford. "Point Estimates of Ordinates of Concave Functions," Journal of the American Statistical Association, Vol. 49, September, 1954, pp. 598–619. (To be reprinted as CCNS, No. 88.)
• Hood, William C., and Koopmans, Tjalling C. (eds.). Studies in Econometric Method, by Cowles Commission Research Staff. Cowles Commission Monograph No. 14. New York: John Wiley & Sons, 1953. 324 pp.
• Houthakker, H.S. "The Specification Problem in Regression Analysis" (abstract), Econometrica, Vol. 21, July, 1953.
• ——. "Electronic Computation in Economic Statistics" (with J.A.C. Brown and S.J. Prais), Journal of the American Statistical Association, Vol. 48, September, 1953, pp. 414–428.
• Hurwicz, Leonid. "Aggregation in Macroeconomic Models" (abstract), Econometrica, Vol. 20, July, 1952, pp. 489–490.
• ——. "Problems of Identifiability" (abstract of discussion), Econometrica, Vol. 20, July, 1952, p. 481.
  Simon, Herbert A. "On the Definition of the Causal Relation," Journal of Philosophy, Vol. 49, July, 1952, pp. 517–528. (Reprinted as CCNS, No. 70.)
• ——. "Spurious"Correlation: A Causal Interpretation," Journal of the American Statistical Association, Vol. 49, September, 1954, pp. 467–479. (To be reprinted as CCNS, No. 89.)
• Telser, Lester G. "Analysis of Variance with a Certain Linear Restriction" (with F.A. Bobkoski). CCDP, Statistics 384.

STAFF CHANGES

Effective July 1, 1952, ROSSON L. CARDWELL assumed the executive directorship of the Cowles Commission, succeeding WILLIAM B. SIMPSON. Simpson continued as secretary of the Econometric Society through September, 1952, and as coeditor of Econometrica until March, 1953, at which time he embarked on a year's tour of the world. Cardwell has had administrative responsibility for the Econometric Society since October, 1952, first as assistant secretary-treasurer and more recently as secretary.

JOHN GURLAND gave up his joint appointment in the Cowles Commission and the University's Committee on Statistics at the end of July, 1952, in order to accept an appointment as Associate Professor of Statistics at Iowa State College.

On August 1, 1952, LEO TORNQVIST, Professor of Statistics at the University of Helsinki, Finland, began a year's appointment as Visiting Professor in the Cowles Commission. His year of research and study was made possible, in part, by a grant provided by the United States Department of State under its educational exchange program. He also received salary support under the ONR contract of the Cowles Commission for his work on the project "Decision-making under Uncertainty."

September, 1952, saw the departure of both WM. L. DUNAWAY and DANIEL WATERMAN. LESTER TELSER, a graduate student in the Department of Economics, replaced Dunaway as a research assistant. Waterman was succeeded as computation leader by EDWIN GOLDSTEIN, who held this position until July, 1953, when he returned to his graduate studies in mathematics at Northwestern University.

STEPHEN G. ALLEN spent the fall quarter, 1952, at the Cowles Commission, on leave from the Applied Mathematics and Statistics Laboratory at Stanford University, to complete his study of linseed oil inventories. He later accepted an appointment as Assistant Professor of Business Administration at the University of Minnesota.

GARY BECKER worked as a research assistant on the Commission's ONR project, starting in October, 1952. He left the Commission the following spring in order to devote full time to his doctoral dissertation in the Department of Economics.

The Commission's RAND project staff, working on "Theory of Resources Allocation," was augmented in mid-October, 1952, by the appointment of CHRISTOPHER B. WINSTEN from Oxford as a research associate. He continued with this project through 1953–54 before returning to the Institute of Statistics, Oxford, England.

At the end of June, 1953, I.N. HERSTEIN moved to Philadelphia to become Assistant Professor of Mathematics at the University of Pennsylvania.

MARC NERLOVE, a graduate student in economics at Johns Hopkins University, joined the Cowles Commission staff as a research assistant for the summer quarter, 1953. When he left in the fall, the portion of his work dealing with transportation was taken over by THOMAS A. GOLDMAN, who continued as research assistant for the balance of the year.

After four and a half years with the Commission, CLIFFORD HILDRETH resigned in September, 1953, to accept an appointment as Professor of Agricultural Economics at North Carolina State College in Raleigh, North Carolina.

JAGNA ZAHL, who succeeded Goldstein as computation leader, left the Commission in early fall, 1953, and was herself succeeded by FRANCIS BOBKOSKI, a graduate student in statistics and a former member of the Commission's computing staff.

ARNOLD C. HARBERGER returned from Johns Hopkins University to the University of Chicago as an Associate Professor of Economics in October, 1953. He later joined the Cowles Commission's staff as a Research Associate, re-establishing a research relationship which began in 1949, when he was a research assistant of the Commission.

In January, 1954, HENDRIK S. HOUTHAKKER began an appointment as Acting Associate Professor of Economics at Stanford University. However, he has continued to direct the Commission's study of markets for future delivery, from Stanford, as a consultant.

RICHARD F. MUTH was appointed Research Associate in the Cowles Commission beginning April 1, 1954, and was assigned to the ONR project "Decision-making under Uncertainty." His appointment succeeded that of ALAN L. TRITTER, whose four-month research assistantship had ended in February.

TJALLING C. KOOPMANS resigned from his position as Director of Research of the Cowles Commission on June 30, 1954, in order to be able to devote more time to research. He has been given a leave of absence for 1954–55 so that he may pursue a program of study and research at Yale University under a grant from the Rockefeller Foundation. ROSSON L. CARDWELL, Executive Director of the Commission, has been appointed Acting Director of Research for 1954–55.

GUESTS AT THE COWLES COMMISSION

COWLES COMMISSION SEMINARS

COWLES COMMISSION PAPERS, NEW SERIES

See complete LIST OF COWLES COMMISSION PAPERS, New Series

     No. 1. JOHN R. MENKE, "Nuclear Fission as a Source of Power," Econometrica, Vol. 15, October, 1947, pp. 314–334.
     No. 3. TJALLING C. KOOPMANS, "Uses of Prices," Proceedings of the Conference on Operations Research in Production and Inventory Control, pp. 1–7. Cleveland: Case Institute of Technology, 1954.

COWLES COMMISSION MONOGRAPHS 1934–1954

See Complete LISTING OF MONOGRAPHS (available for download)

SPECIAL PUBLICATIONS

Commodity Futures: A Study in the Economics of Uncertainty, by H.S. HOUTHAKKER (assisted by LESTER G. TELSER). (Forthcoming.) A study of commodity futures markets which investigates the reaction of groups concerned with these markets to uncertainty about future developments. Contents:

I. Commodity Futures and the Economics of Uncertainty; II. The Futures Contract and Its Uses; III. Theory of Cash and Futures Prices; IV. Interrelations of Stocks, Hedging, and the Basis; V. Gains and Losses of Various Groups of Traders; VI. Concluding Remarks.

Economic Aspects of Atomic Power, An Exploratory Study under the direction of SAM H. SCHURR and JACOB MARSCHAK. Princeton: Princeton University Press, 1950. 289 pages. An analysis of the potential applicability of atomic power in selected industries and its economic effects in both industrialized and underdeveloped areas. Contents:

Preface. PART ONE. ECONOMIC COMPARISONS OF ATOMIC AND CONVENTIONAL POWER: I. Economic Characteristics of Atomic Power; II. The Cost of Electricity from Conventional Energy Sources. PART TWO. ATOMIC POWER IN SELECTED INDUSTRIES: III. The Industry Analyses: A Summary View; IV. Aluminum; V. Chlorine and Caustic Soda; VI. Phosphate Fertilizers; VII. Cement; VIII. Brick; IX. Flat Glass; X. Iron and Steel; XI. Railroad Transportation; XII. Residential Heating. PART THREE. ATOMIC POWER AND ECONOMIC DEVELOPMENT: XII. The Effects of Atomic Power on National or Regional Economies; XIV. Atomic Power and the Industrialization of Backward Areas.

Income, Employment, and the Price Level, notes on class lectures by JACOB MARSCHAK. Autumn, 1948 and 1949. Inquiries should be addressed to the Cowles Commission.