Olav Bjerkholt University of Oslo The Klein Legacy

Olav Bjerkholt (University of Oslo): The Klein Legacy: the Cowles Commission Years How and why Lawrence Klein got to the Cowles Commission. The Tinbergen legacy. The Commission’s style of work. The end of macroeconometric models at Cowles.

Marschak: Research Plans for two years beginning July 1, 1944 (…) It is now planned to revert the facilities of the Cowles Commission to its original and proper field of quantitative economic research. (…) Accordingly, it is intended to develop … an intensive study which may be entitled “Economic Behavior and Business Fluctuations. ” The title suggests our intention to attack the problem of economic dynamics as a whole, by combining any available data on the behavior of both consumers and entrepreneurs, and taking proper account of time-lags and expectations which make the economic system subject to fluctuations. • The work on “Economic Behavior and Business Fluctuations will imply: 1. The setting up of theoretical hypotheses, i. e. of economically “plausible” systems of behavior equations. This will be done partly by eradicating from Tinbergen’s system all those variables which experience and thought will have shown to lack relevance and hence not to justify the expense and further labor. 2. The revision of methodological fundamentals, with particular regard to interdependent systems of time series. 3. The improvement and continuation of some of the time series. The work will be based on the cooperation of T. Koopmans, O. Lange, and J. Marschak. It will require, in addition, assistance of a younger econometrician “at large” (…). Occasional conferences will be held with a few specialists, e. g. T. Haavelmo and Abraham Wald, both currently informed of our work and occasionally advising us.

Marschak’s confidential statement to Social Science Research Council in February 1945 on Lawrence Klein’s application for Fellowship “Before I knew him better, a certain impetuosity of his style in writing and speaking made me expect from him an excess of self-confidence (not unusual in young people of his ability) if not partisanship. On nearer acquaintance this has not been confirmed: my collaborators and myself have found in him a person prepared to understand appreciate the other point of view; equally agreeable in giving and in taking; and more interested in having the problem solved than in winning the argument or making a career. [Klein] has a good eye for the essential. His goal is a logically consistent explanation of observed facts. He will not try to escape into theoretical perfectionism (which tends to make economics logically complete and beautiful but unverifiable) or into empirical detail (substituting enumeration for explanation). The kind of study proposed by Klein requires a sound instinct …. Klein seems to possess the necessary instincts: it will save him much disappointment. Klein seems to be the type of man who will work overnight and over the week-end if the problem interests him. ”

The CC research associate room The entire CC was housed in one huge room except for Marschak, the director, and Dickson Leavens, the business manager, who were in separate offices down the hall. Along the left side of the main room were four cubicles occupied by Koopmans, Leipnik, Rubin, and Lawrence Klein. Along the right side were cubicles for Leonid Hurwicz and me. The central space was used for seminars and had desks for latecomers, one of which was for Haavelmo and another for Sonia Adelson (later to become Mrs. Klein). Because this room faced south it was a pleasant space. The partitions were only seven feet high in a room with twenty foot-high ceilings. However, some of the occupants had strong voices and some liked to talk a lot. The partitions had to be raised to the ceiling. (TWA, Econometric Theory, 2015)

CC staff meeting, Oct. 11, 1946: Specification of money demand. Jacob Marschak (JM), Lawrence Klein (LK), Tjalling Koopmans (TK), Leonid Hurwicz (LH), Trygve Haavelmo (TH), Don Patinkin (DP), Kenneth May (KM) and as guest, John R. Hicks. I. Should securities appear in the individual utility functions in terms of market value or nominal value? LK: Market value. J. R. HICKS and others: Nominal value. TK: If we use market values and the market values of the bonds change, this must be included as capital gains in income. LK: Agreed. J. R. HICKS: Substitute nominal values of bonds multiplied by market prices. LK: The identity holds regardless of the form in which it is written. J. R. HICKS: Allow for both nominal value and liquidity value of securities. JM: The arguments of both the utility function and the production function are parameters of the joint probability distribution of the future yields. LH: LK assumes money itself has a direct utility distinct from the probabilities of income it creates; in his model there is no primary utility function. (…) II. The length of the horizon. (…) III. What variable should be included in the statistical macro equation. (…) [Meeting broken off to continue next morning. ] IV. Supply of money. (…) J. R. Hicks: Money market is not analogous to the commodity market; there can be an equilibrium position with excess reserves.

Staff meeting of December 26, 1944 on equations of production and of labor and equipment demand. (Notes taken by J. Marschak) 1. Marschak remarked on the order of work: a) the three equations for the firm (production equation, the profit-maximization equation with respect to equipment, and the profit-maximization equations with respect to labor; b) summation over firms (“aggregation") results in macro-equations for production, and demand for labor and equipment and involves some knowledge of the distribution of firms by manpower size etc. ; c) completion of the macro-system. Notation: y = output; n = manpower; z = equipment; z 0 = initial equipment … 3. It was agreed that kinks at the “bottlenecks” are less realistic than smooth curves. The choice will therefore depend on relative simplicity in mathematical and other operations. 4. Should the condition that z > z 0 be described by the production function (with a kink or with a vertical tangent at y = o) or should it result from a proper definition of the profit, which would take into consideration all market conditions. 5. Hurwicz: Should one not distinguish between producers and consumers goods? This was postponed to the “macro”- stage of work. … 8. The effect of assuming a given distribution for one of the variables in the aggregative equations. The case of Gibrat distribution discussed.

Haavelmo to Klein November 2, 1945 (with copy to Marschak) in response to Klein to Haavelmo October 31, 1945 and Marschak to Haavelmo October 25, 1945

Three men in a boat – The Ergodic Trygve, Larry, and I bought a used sailboat of modest size. We needed a dinghy (a small row boat) to get from the shore to the mooring. Trygve and Larry were delighted to find a dinghy made by KARL MARX. One fine day Larry and I proceeded to the harbor and took the dinghy to our boat. Neither of us had sailed a boat, but we understood the physics of it. We started raising the sails, but neither of us had tied the dinghy to our boat. To our horror we saw our dinghy being blown by a west wind toward the open water of Lake Michigan. We hurriedly finished putting up the sails and headed after the dinghy. It did not take us long to realize that not only was the dinghy being blown toward Michigan, but we were, too. We gave up the dinghy and looked to save ourselves. We tacked back and forth, but still moved eastward. Finally I went to the bow of our boat and grabbed on to a boat moored near the breakwater. We simply waited for the harbor master in his power boat to pick us up and take us to our mooring. Several days later Trygve told us that he was by the harbor the previous evening and could not see our boat. The three of us proceeded to the harbor. We were dismayed to see that our boat was moored, but the mast was broken and was tied to the hull. We learned that in a storm with an east wind our boat was blown off its mooring and onto the rocks lining the harbor. At that point we gave up our dreams of sailing Lake Michigan. Sonia, Dorothy, and Beulah were very disappointed; they never got on our boat. (TWA, Econometric Theory, 2015)

From left: Ed Boorstein, Dickson Leavens, Lawrence Klein, Gershon Cooper, Tjalling Koopmans, Herman Rubin, George Perazich, Jacob Marschak, Jacques Hartog, Trygve Haavelmo, Sam Schurr, Selma Schweitzer, Gertrude Nissenbaum. Sitting, from left: Don Patinkin, Sonia Klein, Estelle Maas. (March 4, 1947)

Mission accomplished: Economic Fluctuations in the United States 1921 -1941, CC Monograph 11, 1950 Business Cycles in the United States of America 1919 -32 Opening remarks of Chapter 1: This book is written in the spirit of Tinbergen's investigations and is intended as an improvement and extension of his results. As a consequence of the extensive theoretical discussions since 1936…it has become possible to formulate more sharply the structure of the economic system and thereby to gain added simplicity and accuracy not available to Tinbergen at the time of his work. The purpose in building econometric models is to describe the way in which the system actually operates. We want to do more than is suggested by the title to Tinbergen's work, i. e. , more than the mere testing of business-cycle theories. We want also to discover the best possible theory or theories which explain the fluctuations that we observe. If we know the quantitative characteristics of the economic system, we shall be able to forecast with a specified level of probability the course of certain economic magnitudes such as employment, output, or income; and we shall also be able to forecast with a specified level of probability the effect upon the system of various economic policies.

But note the Preface of Monograph 11 “It cannot be overemphasized that this book contains only tentative and preliminary steps in econometric model building. …It is with all modesty that the research results are now presented in published form, and the main hope is that the work contained here may be of some help to other researchers in the field. If I were to begin today the project of reconstructing an econometric model for the United States, I believe that I would proceed somewhat differently, benefiting from knowledge of the weaknesses of the models in this volume. I would expect others to do the same. ”

Aftermath Carl Christ enters The Business Cycle Conference 1949 Spring 1950, Klein gave a seminar at CC. Tentative suggestions of a new Klein model. Klein submitted a sketch of a new model project Reviewed by Tinbergen and Haavelmo. Nothing came out of it.