“Marco Polo imagined answering that the more one was lost in unfamiliar quarters of distant cities, the more one understood the other cities he had crossed to arrive there and he retraced the stages of his journeys, and he came to know the port from which he had set sail, and the familiar places of his youth, and the surroundings of home, and a little square of Venice where he gamboled as a child.”
(Calvino, Invisible Cities)
I. INTRODUCTION
Traveling is one of the oldest forms of knowledge production that exists in almost all cultures. Those who know are those who went away, found out, and came back with something to tell. In discovering the foreign, the traveler does not only acquire new knowledge, but also learns about himself. Traveling is a form of experience that does not reduce to observation, where the subject of knowledge deems itself to be in a privileged and secure situation; instead, the traveler faces, confronts, is exposed to, and is interrogated by, the foreign. He learns whether that which he takes for granted is limited to his own world or can establish a dialogue with the other. It is for this reason that traveling is a primordial historical sujet; in traveling, subjects themselves face what historians must otherwise painstakingly reconstruct: the context.
And so had Tjalling C. Koopmans the occasion to learn about his preconceived notions of knowledge when traveling, with his wife Truus Koopmans, to the Soviet Union in the summer of 1965. Koopmans was the research director of what was the hotbed of mathematical economics in the US, the Cowles Foundation of Research in Economics hosted at Yale University. Cowles represents a profound transformation of the intellectual culture of economics during the decades following World War II from a policy-oriented discipline into a modeling science (Düppe and Weintraub Reference Düppe and Weintraub2014a). Koopmans’s trip led him to the Central Economic–Mathematical Institute (CEMI, also TSEMI), founded two years before in 1963. The CEMI would become the hotbed of research in mathematical economics in the USSR. It would represent a profound transformation of the orthodox Marxist-oriented discipline of political economy into a modeling science that was labeled “economic cybernetics.” Like an ambassador of these two transformations, Koopmans’s travel was carried by his hope of the merging of intellectual cultures of East and West beyond the ideologies that surrounded economic science here and there.
The trip was part of the Inter-Academic Exchange Program between the two Academy of Sciences institutes and run under the auspices of the American Council of Learned Societies (Byrnes Reference Byrnes1969; Engerman Reference Engerman2009, pp. 89ff.). The program had existed since 1958 and was a result of the post-Stalinist liberalization of science. Yet, it was still the Soviet Academy that dictated the fields of research for which they allowed an exchange. Until 1965, they had not called for an economist to come over. Koopmans, who was curious enough to apply, was thus the first official US economist to visit the postwar Soviet Union.Footnote 1
In what follows, I reconstruct Koopmans’s travel using his official report to the American Council of Learned Societies as well as his correspondence with Soviet and US scholars who were relevant to his trip.Footnote 2 Thanks to the increased control of science during the Cold War, we have greater access to the details of the travel than we would have had in more peaceful times. The US officials were curious to know about every aspect of his experience: Whom did he meet? What did they talk about? What impression did they make? The report thus invites us to adopt a first-person perspective focusing on Koopmans’s experience while ignoring other aspects of his trip, of which he might not have been aware.Footnote 3 As partial as Koopmans’s perspective might be concerning the emerging community of mathematical economics in the post-Stalinist Soviet Union, his personal perspective allows us to ask a more profound set of questions: With what hopes and expectations did he leave? Which ideas did he wish to share? Which ideas did he want to bring back? What surprised him? What remained unanswered? And, having gone through these experiences, did he learn something about himself and about the US community of mathematical economists? Thus, reconstructing Koopmans’s trip to the Soviet Union adds a piece to the puzzle of the history of scientific experience in twentieth-century economic science.Footnote 4
Reconstructing Koopmans’s travel from his point of view tells us about the very nature of knowledge produced at the Cowles Commission during the Cold War. It tells us about one of Cowles’s central epistemic virtues, more or less openly adhered to: that of universal reason. If the mathematical innovations at Cowles amount to more than formalist indifference, as the older generation of economists felt at the time, then these innovations should be capable of establishing an exchange with economists in other social contexts who share this same virtue. Facing two opposed political systems, could Koopmans’s theories launch a joint discourse between economists who advance mathematical methods in economics? As Johanna Bockman and Michael Bernstein have argued, Koopmans indeed “believed that cooperation with Soviet and East European researchers could be the foundation of the kinds of confidence-building necessary to forestall World War III” (2008, p. 586). Could it be that mathematical economics overcomes the inherent political tension between a liberal democracy on the one hand and a planned economy on the other? Could it be that Cowles’s new economic science provides common ground from which peace emerges? Or is it limited to the US academic cocoon? In other words, Koopmans’s travel tells us, in his personal terms, of the frictions that are characteristic of Cold War science—frictions between epistemic universalism and political antagonism.Footnote 5
II. OUT OF NEW HAVEN
The fact that Koopmans was the first economist to officially visit the Soviet Union is not surprising. The Soviet sensitivities towards economists were, without a doubt, greater than their sensitivities towards other sciences, which were practiced on more equal terms on both sides of the Iron Curtain. Koopmans, who was socialized as a physicist and mentored by Jan Tinbergen, habitually put his research in the context of market socialism, and thus believed strongly in the equal terms of economic knowledge in capitalism and socialism.Footnote 6 More precisely, he believed in the analytical equivalence of perfect competition and perfect planning such that both systems could be assessed within the same conceptual framework. His entire scientific ethos drew from this attitude that economics should be practiced prior to the specification of institutional terms; and only as such can economics then contribute to the scientific design of economic institutions in whatever system.
From my explorations of Marxist thinking in my student years, I have retained a lifelong interest in the prior formulation of that fundamental part of economic theory that does not require specifying the institutional form of society to be used as a framework for the description and comparison of different economic systems. (Koopmans [Reference Koopmans and Lindbeck1974] 1992, p. 235)
To some extent, this belief in the pre-institutional character of economic knowledge was shared among his US colleagues. How to be a mathematical economist during the early Cold War was not self-evident, considering contradictory social constraints. Many US scientists responded defensively to the ideological pressures inherent in the Cold War academic milieu, exerted by phenomena such as military secrecy, McCarthyism, but also anti-Semitism, as E. Roy Weintraub has recently argued for economics (2014; see also Düppe Reference Düppe, Pahl and Sparsam2013).Footnote 7 The rhetoric of depoliticized research appealed to many scholars averse to power politics, specifically those who applied methods that stem from military research, which was easily confused with a non-democratic use of science. Also Koopmans contributed to military war planning as a statistician for the British Merchant Shipping Mission in Washington by analyzing efficiency problems of the transportation of tanks (1939, 1942). It was in this research that he formulated the principles of what he would, after the war, call “activity analysis” (Koopmans 1951; Düppe and Weintraub Reference Düppe and Weintraub2014a).
Applying to the Soviet exchange program, Koopmans showed that his notion of pre-institutional economic knowledge was more than a mere rhetorical strategy. Since Cowles’s mathematical economics was hardly ever explicitly confronted in an open philosophical discourse (Düppe Reference Düppe2011), it was unclear if it amounted to more than just formalist indifference, but rather to actual universal theories valid across the political systems once applied. If economic theory is pre-institutional, then it should play the same role in a free market economy and a socialist economy. Compared with many of his colleagues, who would easily sneer at Soviet economists, Koopmans was thus the ideal candidate for approaching Soviet economists on eye-level.Footnote 8 He operated simultaneously in what for others were two separate spheres: universal reason and institutional design. Traveling to the Soviet Union, Koopmans had the opportunity to put the new regime of economics at Cowles to test: Was it capable of trans-political travel, or was it, as many suspected, an inward-looking practice prospering only in the cocoon of the Cowles Commission? Thus, Koopmans left New Haven with the hope of creating a trans-political community of mathematical economists—economists who could help one another optimize the systems of their own countries.
Koopmans had good reasons for his hope. After the Stalinist repression of science came to an end, a seemingly similar belief in pre-institutional, scientific foundations of economic policies also became fashionable in the Soviet Union. In the same years that the wall was built in Berlin, reconnaissance aircrafts were sent to Cuba, and the war escalated in Vietnam, scientific optimism about reforming economic planning took off in the Soviet Union under the label of economic cybernetics. In 1958, the Academy of Sciences’s Laboratory of Economic–Mathematical Methods was established thanks to the initiative of Vasily S. Nemchinov (1894–1964). In 1959, two book publications marked this turn: Nemchinov gathered five essays by himself, Oskar Lange, Viktor V. Novozhilov, Leonid V. Kantorovich, and Alexander L. Lur’ie; and Kantorovich published his long-awaited “pamphlet” on the efficient use of resources. The subsequent debate among economists took place at a conference in April 1960 on the use of mathematical methods in economics (Campbell Reference Campbell1961; Ellman Reference Ellman1973). As a result, Nemchinov became a chairman of the mathematical economics Council of the Academy of Sciences. In 1961, cybernetics was made the official party line at the 22nd Congress of the Communist Party. At this point, Soviet economic cybernetics developed a distinct culture independent from what Norbert Wiener would be able to recognize. All fields of research that were in some way related to the idea of automated planning were placed under this new label (which included far more fields of research than in the West: operations research, game theory, and mathematical linguistics, among others).Footnote 9 Even the official Philosophical Section of the Academy of Sciences reinterpreted dialectical materialism in terms of cybernetics such that “information” could not spoil “matter.” The main supporters of economic cybernetics were Nemchinov, Kantorovich (1912–1986) and Novozhilov (1892–1972).Footnote 10 In 1965, they received jointly a Lenin prize, the highest academic honor in the Soviet Union. Koopmans would not meet Nemchinov, as he died shortly before his visit and shortly after he founded CEMI (Medow Reference Medow1965).
The economic cybernetics boom happened against the backdrop of the resistance of the Marxist economic orthodoxy, such as Stanislaw Strumilin and Adolf I. Kats (Katsenelinboigen Reference Katsenelinboigen1980, pp. 140ff.). This orthodoxy refused the formal equivalence of capitalism and socialism, considered any diversion from the labor theory of value as bourgeois, thought of the plan as an adaptive “living” plan, and was afraid of losing power if its “living bureaucratic labor” was subverted by a machinated plan (Sutela Reference Sutela, Barnett and Zweynert2008). Koopmans’s invitation by the Soviet Academy of Sciences was a clear sign that this orthodoxy had lost influence. He would clearly not be allowed to meet, and neither was he interested in meeting, orthodox Marxists. He certainly would not want to actively participate in their criticism of mathematical economics, though he knew well of its existence.Footnote 11
Compared with the economics of the Stalin period, scientific discussion of economic problems in the USSR has gone a long way in freeing itself from the shackles of doctrine. One can infer that the highest authorities now support a systematic scientific effort to bring tools, ideas and manpower from mathematics, mathematical programming, cybernetics, economics and statistics to bear on the problems of economic planning. (Report)
Even if the cybernetics hype was carried to a large extent by futuristic sentiments, it was paralleled by an actual reform of economic planning. The months before and after Koopmans’s visit, the public debates preceding the so-called Liberman reform of September 1965 were at their peak. The Liberman reforms granted more independence to firms in reducing the amount of their objectives. Firms would employ their own funds, which, one hoped, would induce growth by increased efficiency (Sutela Reference Sutela1991, pp. 54ff.). It was exactly this kind of reform that would suggest using a model very similar to Koopmans’s own model of production. Koopmans would, as we see below, discuss with most economists “the choice of objective function of planning,” though never directly about the still rather contested Liberman reforms.
Koopmans’s curiosity about Soviet economists was not only caused by the cybernetics hype in the years preceding his visit. Eight years before his travel, he learned of the work of Leonid Kantorovich on linear programming, which he considered equivalent to his own model of activity analysis ([1939] Reference Kantorovich1960, [1942] Reference Kantorovich1958, [Reference Kantorovich1959] 1962). Prior to this discovery, Koopmans considered himself to be in the first row, jointly with George Dantzig, of the pioneers of linear programming. His model stemmed from WWII military research and informed much of the emerging research after the war, including general equilibrium theory, growth theory, computational methods, and also game theory. The activity analysis model was a theory of optimal production in which inputs and outputs are in fixed relations and subject to a set of inequalities. Specifically delicate was the interpretation of the Lagrange multiplier in terms of marginal costs, which could be, but did not have to be, interpreted as market prices. Koopmans himself was uncertain as to how to interpret the multiplier, and wavered between the concepts of “values” and “prices.” He finally called them “shadow prices” (Düppe and Weintraub Reference Düppe and Weintraub2014a). It was clear to Koopmans that his model could be used as a tool for production planning with shadow prices as the objective variable, even though he thought of “planning” as an organizational necessity rather than a political option.Footnote 12
Koopmans learned of Kantorovich’s work from Merill Flood, who in turn knew it from the mathematician Max Shiffman (Koopmans Reference Koopmans1960, p. 363). In November 1956, with the help of Raymond Powell, he read Kantorovich (1940) and immediately contacted him:
you have in part paralleled but in greater part anticipated a development of transportation theory in the United States which has stretched out over the period from 1941 to the present and is still continuing.… Your brief article contains in beautiful summary the mathematical essence of what was developed here. (TKP 13, “Kantorovich, Leonid V., 1956–59”)
Koopmans was excited because despite the apparent different political pressures, the resulting theoretical innovations were the same. Learning about Kantorovich, for Koopmans, corroborated the generality of his theory. Hence, he put considerable effort into the diffusion of Kantorovich’s work (Koopmans Reference Koopmans1959, p. 68n; 1960; 1962). However, the correspondence with Kantorovich was sparse, and the negotiations with his US colleagues were spiked with prejudices. Not everybody was ready to accept Kantorovich’s priority. Koopmans noted an element of “national pride” when rebutting the critique of Abraham Charnes and William Cooper (Reference Charnes and Cooper1962, p. 265).Footnote 13 Traveling to the Soviet Union, Koopmans was looking for more corroboration of his style of theorizing beyond nationalist sentiments, which were so widespread at the time.
Next to the discovery of Kantorovich’s work, Koopmans had also received other signs that spurred his curiosity about mathematical economics in the East. In November 1961, John Montias from the Cowles Foundation visited Czechoslovakia and Hungary, and reported enthusiastically to Koopmans:
One young man at the Vysoka Skola Ekonomicka in Prague recently wrote his ‘diploma thesis’ on a non-linear inter-industry model, a byproduct of which was what he claims to be a new solution to concave programming problems; in Budapest, work will be started soon on a large-scale L.P. problem where the inflow of raw material into the system is treated stochastically. (TKP, carton 22, “Trip to the Soviet Union”)
Stochastic use of linear programming was just what was happening on the front line of research in the US, beginning with Dantzig (Reference Dantzig1955). Not only did the methods appear to be the same, but also the aim. Mathematical economists in the Soviet Union were moved by the same dream that Koopmans had for his entire life: economic efficiency.Footnote 14 “Kantorovich has estimated,” Koopmans enthused in his report, “that better planning methods alone could raise output of finished products by 30 to 50 percent.”
Koopmans was not the only US scholar interested in Soviet economics. In the early 1950s, the Ford Foundation established the so-called Area Studies on the Soviet Union in various disciplines, although initially few economists were involved. The RAND Corporation was leading in fostering research about Soviet economics. Since the early 1950s, Norman Kaplan, Alexander Gerschenkron, and Abram Bergson were hired to get a better grasp on the Soviet economy (Engerman Reference Engerman2009, p. 105).Footnote 15 Since 1950, Wassily Leontief brought together scholars from the East and the West in an annual conference series held in western Europe (Bockmann Reference Bockman2011, p. 65). The fruitless attempt of Jacob Marschak, in 1947, to contact the Soviet Academy of Sciences for the invitation of Soviet economists demonstrates how difficult it was to establish contacts during the Stalinist period. In 1961, interest in Soviet planning also began to arise in Norway when, with the support of Ragnar Frisch, the Journal of Economics of Planning was launched in the light of perceived merging problems of economic policy in the East and West.Footnote 16 Just the day before Koopmans took the plane, between May 7 and 9, a Conference on Mathematical Techniques and Soviet Planning was held at the University of Rochester, which brought together Sovietologists and economists.Footnote 17 It was from this publication that the field of “comparative economics systems” took root; this field would remain part of the US economics curriculum until the 1980s. Koopmans’s travel, as we will see, is revealing regarding the nature of interest in this field.
There is no evidence that Koopmans knew about US–Soviet Studies, which were indeed of little importance to his goal of finding like-minded scholars with whom to work on Cowles’s theories. He was less interested in Soviet planning practices per se, but curious as to how his theory could be applied in the context of planning. Thus, Koopmans went to the Soviet Union with one big question in mind: Is linear programming used as a planning tool in order to determine what he called “shadow prices” and Kantorovich called “objectively determined valuations”? Does his theory of production inform the Soviet centralized allocation of resources, leading to greater efficiency?
I pressed and had persistently the question whether the mathematical methods were to be used to perfect centralized planning (through the use of Kantorovich’s “objectively determined valuations”, “shadow prices” is a western term), or to motivate and guide the use of more centralized allocation through markets. (Report)
Reinforced by certain cautiousness from his side and secrecy from the other side, he would not learn much about the principles that guide actual planning at GOSPLAN. He would only inquire into the theories of envisioned planning, into perfect planning. Rather than learning from Koopmans about planning methods of the USSR, the question for us is: Given his expectations, hopes, and questions described in this section, which kind of meetings would Koopmans profit from and enjoy, and which would be rather disappointing?
His itinerary was packed: four days in Warsaw, ten days in Moscow, five days in Leningrad, a stopover in Moscow to fly to Novosibirsk for three days only, to return after another stopover in Moscow to fly to Budapest and then back to Yale. He would meet and talk to around sixty different scholars of considerable variety.Footnote 18 Which were the most telling meetings?
III. MOSCOW
On May 10, Koopmans arrived at the Central Economic–Mathematical Institute (CEMI). The institute was founded two years before in 1963 and was the immediate manifestation of the economic cybernetics boom described above. Before that, there would have been no institution that could have hosted Koopmans. The institution was large-sized, comparable to institutions such as RAND or the Bureau of Labor Statistics rather than Cowles. Three hundred researchers, he was told, worked for CEMI in Moscow alone, while two other “branches” existed in Tallinn and Leningrad.
The institute was created in order to carry out its core project for which existing institutions lacked the infrastructure: a general theory and the implementation of a single, automated, nationwide system of economic control. The project was called SOFE—the System of Optimal Functioning of the Economy (Sutela Reference Sutela1991, pp. 37ff.). Two vital elements of this system were a mathematical model of the entire Soviet economy, and a central nationwide network of information that is sometimes referred to as the failed “Soviet Internet.” Planning should be optimized by a decomposition of tasks on several levels of the economy, the coordination of which required some notion of prices. The practical aim of the project was to provide continuous day-to-day planning on the basis of a “long-range” (fifteen to twenty years) growth plan that was to build up a communist society (Zauberman Reference Zauberman1975). An early manifestation of this project was a substantial investment in the coal industry (growth was traditionally seen as a matter of choosing the right industry in which to invest). Nikolay P. Fedorenko (1917–2006), who had previously been at the Institute of Cybernetics, became the head of the project and, thus, the first director of CEMI. He was also the first person Koopmans would meet upon his arrival.
In Koopmans’s eyes, Fedorenko must have been an unlikely director. He was a chemist without experience in computing technology, who did not begin working on national economic modeling until he became the director of CEMI. Even then, his main research activity was the management of chemical industries (Kassel Reference Kassel1971). Fedorenko introduced Koopmans to the five departments of CEMI—roughly said, planning, theory, control, mathematics, and hardware—which were subdivided into further “sectors” and a total of around fifty “laboratories” only in Moscow.Footnote 19 They also spoke about the institute’s research agenda as set in 1964:
(1) Elaboration of a theory of optimal planning and management, and the construction of a general mathematical model of the national economy
(2) Development of a unified system of economic information
(3) Development of a unified state network of computation centers
(4) Development of mathematical methods for the general model
(5) Creation of concrete planning and management systems based on mathematical methods and computer technology
(6) Elaboration of standards and algorithms for planning and management. (Gerovitch Reference Gerovitch2002, p. 373; see also Kassel Reference Kassel1971, p. 94n)
To Koopmans, this agenda might have looked similar to that of the research department of GOSPLAN, which had the primary responsibility for economic planning. The CEMI, an Academy of Sciences Institute, however, was less tied to the party line than was GOSPLAN. As Olessia Kirtchik and Ivan Boldyrev argued (2014), CEMI was indeed allowed to conduct independent research free from the imperative of immediate relevance, specifically in its mathematical section. The relationship between GOSPLAN and CEMI was not obvious: data-intensive input-output analysis here, and theoretical visions of centralized networks there; piecemeal ad hoc planning here, futuristic science of full control there. The CEMI was created with respect to a scientific vision that, to a large extent, bypassed current planning practices.Footnote 20
Koopmans also learned of a new journal edited by Fedorenko, Economics and Mathematical Methods. He was given copies of the first two issues. He understood that CEMI staff did study the Western literature very carefully, and translate books and papers. They had access to a large range of Western periodicals, including Econometrica. Fedorenko understandably tried to show that his institute could compete with Western institutes. Koopmans, instead, had his “pressing question” in mind. Perhaps he had read Fedorenko’s first annual research report, which may have made him hope for a vivid conversation. There, Fedorenko drew an impressive image of the first planning successes: “The institute conducts work on the creation of methods of optimal planning and management of transportation in the country. Over 1000 optimal plans … have been computed. The savings of this work have already reached about half a billion rubles” (in Kassel Reference Kassel1971, p. 94). Yet, the first answer he got from Fedorenko must have been disappointing. Fedorenko showed him a map of the distribution of chemical plants to which he contributed without giving deeper insights into the functioning of the system—perhaps he could not, perhaps he wished not, perhaps he was not supposed to give him this insight. Koopmans could only speculate. Fedorenko was one of those about whom Koopmans would say later that he was kept at arm’s length. His representative position, however, was enough for Koopmans to suggest him as a new member of the Econometric Society in 1966. Other, more mathematically minded Soviet economists would have to wait longer for this honor.
During the coming days, Koopmans met several heads of the CEMI sections that interested him. The first notable meeting was with the young Evgenii Gol’shtein of the mathematical programming section.Footnote 21 Gol’shtein was in his early thirties, and on his way to become the leading researcher on computational mathematics and programming: Block programming, duality theory, modified Lagrange functions, non-differentiable convex optimization—very much Koopmans’s sort of things. Gol’shtein knew Koopmans’s work and would also extend it. The overlap between their work impressed Koopmans, so much that he would call Gol’shtein, after Kantorovich, the second-best programmer in the Soviet Union, which likely had an impact on Gol’shtein’s steep career. Koopmans also met Aron Katsenelinboigen (1927–2005) (with collaborators Boris N. Michaelievski, Victor A. Volkonski, Vladimir A. Masch, and Alexander L. Lur’ie). Katsenelinboigen was the head of the Department of Complex Systems that later became the department of mathematical economics, which employed the most mathematicians and also the most Jews who could not find jobs elsewhere.Footnote 22 Koopmans spoke with Katsenelinboigen about “the choice of objective functions of planning,” which comes down to the problem of either using the price mechanism or giving direct orders. Koopmans also met Anatoly Modin (1935–1994), who was already known to Jacob Marschak. Marschak was curious about Koopmans’s impressions:
I found the conversation with Modin somewhat disappointing. I think he is trained as an engineer, and the conversation seemed to go in terms of writing down box-arrow-diagrams showing how information flows between different levels. Language difficulties did not make the conversation any easier. (Koopmans to Marschak, October 27, 1965)
In fact, language was one of the obstacles that created a natural distance in most of the meetings. Koopmans initially planned to learn some Russian, but did not get beyond the basics. His Soviet colleagues, specifically the young, instead studied English. Meeting Koopmans was for many the first occasion to talk to a foreigner in another language (Makarov, personal conversation). For most conversations, Koopmans had a translator, Mr. Altaev, an engineer at CEMI, who was at the same time a guide traveling with Koopmans to both Leningrad and Novosibirsk.
After the weekend, on May 17, Koopmans gave his first talk at CEMI. The projector slides, which were of US standard size, turned out to be too large for the overhead projectors from eastern Germany. The choice of talks was not obvious. Which topics would not be considered offensive, which useful, and which intelligible?
In choosing topics for lectures, the main choice open to me was between a) theorems on the efficiency of competitive market allocation and b) presentation of recent work by western economists on models of optimal growth. I chose the latter in order to stay away from the area of suspected dispute.… Judging by facial expressions and subsequent comments, they were well received. I was agreeably surprised by the almost total absence of “doctrinal” debate in the often extended discussion periods. (Report)
The four talks, given on about fifteen occasions at universities and Academy of Sciences institutes of both mathematics and economics, were, first, “an introduction to the von Neumann model of proportional growth”; second, “the turnpike theorem on maximal non-proportional growth”; third, “optimal economic growth paths for an indefinite future”; and fourth, “on the existence of subinvariant measure and an application to economics.”Footnote 23 Due to John von Neumann’s reputation among mathematicians, everyone in the Soviet Union working with an interest in cybernetics and, specifically, the theory of automata knew his work (Makarov, personal conversation). Valery Makarov had already published a paper, in 1962, on von Neumann’s existence proof (see also Morton and Zauberman Reference Morton and Zauberman1969). Therefore, it was not Koopmans who imported von Neumann’s growth model to the Soviet Union; however, he would later claim that his two presentations on growth theory did make a difference to Soviet economics. His article on growth (1965) was quite successful in the West, later labeled the Ramsey–Cass–Koopmans model. Optimal growth theory in the Soviet Union—which led away from the belief that one only has to pick the right industry to invest in and then wait for the spillover—might have been a Koopmans import:
I believe I have been effective in the lecturing I did in 1965 describing to Soviet mathematical economists work in optimal growth theory done in the West. This was confirmed by Dr. Gvishiani, who stated to me in 1974 that I had an intellectual influence in the U.S.S.R. I made these efforts, that apparently had a result, because I think it is important to strengthen rational economic thinking in the second most powerful nation in the world. (Koopmans to Kaysen, February 8, 1977)Footnote 24
Note that in this remark, made more than ten years after the trip, Koopmans presents himself less as an ambassador of universal reason, but rather, with a certain sense of superiority, as a missionary of local reason. There is no evidence of such an attitude in his earlier reports.
Koopmans not only exported but also imported knowledge from the Soviet Union. But this import did not take place via economists but rather via mathematicians. Mathematical culture was, much in contrast to that of the US, deeply ingrained in Russian society and one of the strongest pillars in academia also during the Soviet period. In order to be received in audience by mathematicians, prior to his trip, Koopmans went to the Courant Institute of Mathematical Sciences at New York University to ask for help from Jack Schwartz and Richard Courant. But it was CEMI that had to arrange the meetings in the last instance. Koopmans sensed that they resisted doing so.
In Moscow my appointments and travel schedules were all arranged through a higher staff member of the CEMI, Dr. Gorfan. I sensed some correlation between the ease and success in getting an appointment or travel arrangement and my estimate of CEMI interest in such appointment coming about. However, the foreign visitor who is held at arm’s length may also mistake the effects of randomness and/or sloppiness in organization for deliberate intent. (Report)
Although he did not get all the meetings that he wanted, he nevertheless managed to meet the crème de la crème of Russian mathematicians. He spoke to one of the most well known at the time, Andrei Kolmogorov (1903–1987), who worked on the axiomatization of probability theory. Kolmogorov, Koopmans writes, was well informed about mathematical economics at CEMI. He met Aleksei Lyapunov (1911–1973) (not to be confused with Alexandre Lyapunov), the founding father of cybernetics in Russia. He also met Andrei Tikhonov (1906–1993), an important topologist. But most importantly, thanks to Richard Courant, he met Israel Gelfand (1913–2009), a student of Kolmogorov known in the East and the West. The three meetings with Gelfand and his collaborators could have led to an actual import of ideas to the US:
Dear Professor Gelfand, I am writing to thank you again for the time you and Prof. Zetlin have given me for what I have found to be extremely valuable discussion. In particular I think that games with simple-minded players will turn out to be an extremely fruitful idea in the social sciences (I cannot assess their value in biology, but I take your word for it that the idea is valuable there too). (Koopmans to Gelfand, June 2, 1965)Footnote 25
Back home, Gelfand’s work would be the only research that Koopmans wished to share with his colleagues. In October 1965, he would send a list of references (including, for example, Gelfand et al. Reference Gelfand, Piatelki-Schapiro and Zetlin1963) to the leading game theorists in the US: Robert Aumann, Gérard Debreu, John Harsanyi, Oskar Morgenstern, Jacob Marschak, Anatol Rapoport, Thomas Schelling, Herbert Scarf, Lloyd Shapley, Martin Shubik, Herbert Simon, Robert Strotz, Robert Thrall, Albert Tucker, and Harold Kuhn. Koopmans explained:
My purpose is to report briefly on conversations I had in Moscow last May with I. M. Gelfand … and M. L. Tsetlin … in constructing a theory of games in which the players are extremely limited, in memory observation, and analytic powers—the opposite extreme to von Neumann Morgenstern game theory. There appears to be an extensive literature in this field in Russia … one of the motivations arose from looking at the operation of the nervous system as an interaction of its cells. While the “behavior” of each cell is presumably simple, the performance of the entire nervous system is of a much higher order of complexity and effectiveness. I also have an impression that work along these lines has a potential bearing on economics as well. Not all of us are perfect calculators in our economic decisions. The analysis seems to be a search for cases or conditions in which the simple players unwittingly achieve behavior that is asymptotically efficient or optimal under repeated play.Footnote 26 (TKP, carton 22)
Evolutionary game theory could have been a Soviet import! However, hardly anyone was really interested, as Koopmans reported to Gelfand on November 3, 1965:
My experience so far has been that, the more the individual in question is actively doing mathematical research in game theory, the less he is inclined to stop and contemplate quite different formalizations of games.... In closing, allow me to say that I found our conversations among the most stimulating I have had in my month in the USSR.
Koopmans tried to further promote Gelfand by asking Franklin M. Fisher, the Econometrica editor, to ask Gelfand to write a survey article of the Russian literature on simple-minded games. It did not happen. As many scholars in the potential research community across the Iron Curtain suffered from the political divide, so too were there others who profited from the divide, as Koopmans had to learn about his colleagues.
IV. LENINGRAD
On Friday, May 21, Koopmans traveled to Leningrad, where another branch of CEMI, smaller than that in Moscow, has just opened. There, the atmosphere was less tense than in Moscow, since there were no other institutions from which CEMI officials had to keep Koopmans away. He emphasized how friendly his reception was in Leningrad. The institute emerged from the Computing Center of the prestigious Leningrad branch of Steklov’s Mathematical Institute (LOMI). The departmental structure of CEMI Leningrad consisted of six laboratories; valuation, game theory, and queuing theory were the most prominent. The central laboratory of valuation was run by one of the three main protagonists of economic cybernetics, the seventy-three-year-old Viktor V. Novozhilov.
Novozhilov was a Marxist economist with mathematical ambitions and knowledge of neoclassical doctrines—a figure comparable to Eugene E. Slutsky, or even Vladimir Dmitriev (see Barnett Reference Barnett2005). He worked for most of his life on a theory of measuring costs as the key to economic planning, which was also central to Koopmans’s activity analysis model. Novozhilov always tried to reconcile his theory with the labor theory of value, but always acknowledged the relevance of the price mechanism (see Campbell Reference Campbell1961; Hagendorf Reference Hagendorf2012). Koopmans could have read his contribution to Nemchinov’s collection of essays (1959), translated in 1964, as well as a French translation of his work (Novozhilov Reference Novozhilov, Novozhilov, Johansen, Chambre, Fiszel, Kreyberg, Waelbroeck and Stolz1964). But, apparently, Novozhilov did not impress him, as he did not mention more than his name in his report. Koopmans might have already given up pressing his question very hard with senior economists and preferred instead to establish friendly relationships by talking about shared passions of a different kind, such as music—Koopmans had a lifelong interest in classical and modern music, and also tried himself as a composer.
I am most grateful to you for … your sending me the record of Beethoven’s trios minus one violin.… On my turn I am sending you the record of Glinka’s opero “Ruslan and Lyudmila.”… Your book “Three Essays …” and the article “On the concept of optimal Economic Growth” are wonderful with their deep thoughts and brilliant form. (Novozhilov to Koopmans, November 24, 1967)
The shared esthetic appeal of art, apparently resembled in science, can establish more easily a common sphere among scholars who do not share the same political commitments.
One might expect that the meeting with forty-year-old Nikolai Vorob’ev (1925–1995), who is considered the founder of Soviet game theory, was more substantial. He was a student of the mathematician Andrey A. Markov (known as the creator of Russian constructive mathematical logic); he then turned to probability theory, and then to game theory, specifically coalitional games. In 1955, he published the first paper in game theory in the Soviet Union (“Controlled Processes and Game Theory”); his second on bimatrix games (1958) was later improved by Harold Kuhn, and a third survey article (1959) opened the door for generations of Russian scholars to venture into Western game theory. In 1968, he would translate von Neumann’s and Oskar Morgenstern’s book. At the time of Koopmans’s visit, Vorob’ev worked on a generalization of Harold Kuhn’s equivalence theorem (of mixed and behavioral strategies). All in all, the game theory that Koopmans could discuss with Vorob’ev was clearly a US import. The conversation might not have been very different from those that he could have with his colleagues at home. Perhaps for this reason, Vorob’ev is barely mentioned in Koopmans’s report —there was little to learn from him, specifically after the impressive discovery of Gelfand’s work.
After meeting several other researchers, on May 26, he and his wife went sightseeing in the company of Alexander A. Korbut and Joseph V. Romanovsky. They drove twenty miles west of the town to see the castle of Petrodvoretz, the Russian version of Versaille. Koopmans might have enjoyed the conversation with Romanovsky on dynamical programming as much as the castle, as he counted Romanovsky among the best minds he met. However, the little tour on the countryside was rushed because they had to be back in the afternoon in order to discuss the teaching program in mathematical economics at Leningrad University. The program was split between the mathematics and the economics departments, while the students were mainly math students—a recruitment dilemma that was also true at Cowles. Throughout the 1950s and even 1960s, mathematical economists were hardly ever recruited from economics departments.
V. NOVOSIBIRSK
On June 5, after a short stopover in Moscow, Koopmans flew 2,000 miles to Novosibirsk to the Siberian Branch of the Academy of Science, founded in 1957. Twenty miles south of the city, a vast “Academy Town” (Akademgorodok) has been created that hosted around 50,000 inhabitants. There, at the Laboratory on the Application of Mathematics in Economics of the Institute of Mathematics, not part of CEMI, he finally got acquainted with Kantorovich.
Kantorovich, two heads smaller than the tall Dutchman, was already a symbol of the post-Stalinist mobilization of science.Footnote 27 Trained as a functional analyst, in the late 1930s he was to improve production processes in the plywood industry, likely commissioned by Nikolai S. Voznesensky, GOSPLAN chief since 1937. Having thus formulated a first version of what was later called “linear programming” (1939, 1942), he saw its potential for national planning. He submitted his proposals twice to GOSPLAN, in 1942 and in the mid-1950s, but without success. In 1949, despite being a Jew during the height of Soviet anti-Semitism, he received the Stalin prize for his work in mathematics. But it was only after Stalin’s death that his ideas of improved economic planning could be published to reach a broader audience. His book, The Economic Calculation of the Best Use of Resources, written in 1942 and published in 1959, caused heated discussions about the use of mathematical methods in planning with respect to the official labor theory of value (such as in Kats’s review of 1960)—and this despite the fact that Kantorovich was very cautious in his ‘economic’ interpretations. Anticipating the critique that mathematical methods could be seen as apologetic of Western economics “forced me … to avoid the term ‘economic’ as much as possible when talking about the organization and planning of production; the role and meaning of the Lagrange multipliers [the “shadow prices”] had to be given somewhere in the outskirts of the second appendix and in the semi Aesopian language” (Kantorovich Reference Kantorovich and Leifman1990, p. 31).Footnote 28 Koopmans, when propagating his equivalent theory in the US, had a similar problem, since he also had to downplay the meaning of prices as understood by market socialists (Düppe and Weintraub Reference Düppe and Weintraub2014a).
At other moments, Kantorovich was less cautious, and rather outspoken in his belief in the superiority of his methods:
The computer cannot digest some of our economist’s scholarly products.… Any attempt to give them a logical-mathematical, algorithmic form in order to enter them into a computer failed. It turned out that after removing everything that was said “in general” … and after pouring out all the “water”, there was either nothing left, or just one big question mark, the formulation of an unsolved problem. (Kantorovich Reference Kantorovich1959, in Gerovitch Reference Gerovitch2002, p. 276)
Apart from such rebellious moments, Kantorovich mostly avoided confrontation. Atypical of an academician in the USSR, Kantorovich was in good contact with influential GOSPLAN members such as Nikolai K. Baibakov and Veniamin E. Dymschitz (Makarov, personal conversation). Such contacts required a certain political conformism, which is emphasized by Katsenelinboigen (Reference Katsenelinboigen1978). Kantorovich would not have been willing to declare in public that the labor theory of value contradicts his ideas. The first report that Koopmans received about the intellectual stricture of his soulmate presented a picture of a rather timid, if not intimidated, scholar. In 1957, Raymond P. Powell reported from his meeting with Kantorovich:
I did succeed in seeing Kantorovich in Leningrad.… The interview was the oddest of my experiences in the Soviet Union.… The peculiar aspect of the exchange was that Kantorovich appeared almost paralyzed with nervousness. He barely spoke about a whisper. His hands trembled markedly. If sweat did not break out of his forehead, it looked as though it should. He insisted on speaking English, though his English was little better than my Russian.… He knew of ‘some’ practical applications of his ideas within single plants, though he spoke only of his fright-flow proposals. When I pressed him on the extent of such applications … he shrugged his shoulders, Perhaps! Had the State planning Commission shown any interest in his ideas? “Well, they were in print; the people at GOSPLAN could read them if they wanted to.” The impression which he very clearly conveyed … was that of a man who disclaimed all responsibility for, and of interest in, his ideas, once they had left his pen. (Powell to Koopmans, August 1957)
Koopmans would repeat this impression of a reticent scholar also years later after the joint Nobel Prize. “His contributions are somewhat concealed by self-imposed political cautiousness in the style of writing, sometimes I think beyond the call of duty and necessity” (Koopmans to Kaysen, February 8, 1977). Anticipating Kantorovich’s sensitivity, it might well be that Koopmans remained cautious regarding his most pressing question. According to Makarov, they always used an interpreter when talking to one another, even at dinner at Kantorovich’s home. Whatever impression Kantorovich made, Koopmans was convinced that he was the outstanding scholar in the USSR. Upon coming home, he would distribute a list of references of Kantorovich’s writings on “optimal planning” to a row of colleagues—although, hardly anyone could read Russian.
During the second day in Novosibirsk, Koopmans gave a survey lecture to an audience of 200 people. It was the final lecture of a conference on mathematical programming. Many in the audience were rather young scientists open to new ideas. The new academic town in Novosibirsk specifically attracted young scientists to launch their careers, as the older generation was less mobile. Far away from the center of power, a new generation of scientists grew up that had never experienced the Stalinist regime of science, which their teachers’ generations still vividly remembered. It was with them that Koopmans felt more at ease talking. He met, for example, Abel G. Aganbegian, later Mikhail Gorbachev’s adviser, who, at the time, was thirty-two years old and a rising star in science management. Also Kantorovich’s assistant, the twenty-eight-year-old Valery L. Makarov, had a young and motivated mind. Makarov, the president of CEMI today, would have a fast career in Novosibirsk, the Academy of Science, and on the international scene. Koopmans was impressed by Makarov and mentioned him among the first-rate minds (letter to Kaysen, 1977). He was one of the few who was trained as an economist and only self-trained as a mathematician. Koopmans could talk with him about general equilibrium theory. Makarov was the first in the Soviet Union to publish, in Reference Makarov1962, on this Cowles specialty. Makarov and many other mathematical economists did not perceive any contradiction between general equilibrium theory and the institutions of economic planning. This was not because general equilibrium theory was considered a theory of planning, as Oskar Lange would have it. Instead, abstract mathematical economics found support and interest because of a certain eclectic approach to economic methods. There was little dogmatism regarding the possible sources of inspiration for improved planning, which made possible many contributions by Soviet scholars, starting from Makarov, to general equilibrium theory (see Kirtchik and Boldyrev Reference Kirtchik and Boldyrev2014). The socialist dreams in Walrasian economics were more visible when Leonid Hurwicz presented mechanism design to a Western audience than when Makarov presented an existence proof without the assumption of insatiability to an Eastern audience (Reference Makarov1965).
Koopmans’s assessment of the situation of the young Soviet mathematical economists is worth quoting. He identified the two main forces that also impinged upon mathematics in the US Cold War regime:
Among the younger people there is somewhat of a tendency toward abstraction as an end in itself, of which we have seen so much more in the West. The fact that Russian mathematics is more tolerant and supportive toward application than Western mathematics helps limit this tendency. On the other hand, the risk in addressing really important economic problems which the Soviet system has difficulty coping with works in the other direction, particularly for the younger people. (Report)
These two forces, the one pushing forward into application, the other pulling backward into abstraction, were the very same forces present during the 1950s in the US when military and industry demands were mingled with the social angst that reached its peak with McCarthyism. The social pressures that the mathematicians had to deal with, specifically the younger generation, resulted from the same “image conflict” between the pure and the applied, that Amy Dalmedico has observed for the US (2001). At last, Koopmans must have had an idea of the symmetry of how the Cold War conditioned the practices of science—an idea that could have helped him better understand the historical specificity of his epistemic virtue of universal reason.Footnote 29
VI. BACK HOME
Back home at Yale, after a total of six weeks of traveling, how did Tjalling and Truus Koopmans feel? One of the familiar things for which they were certainly grateful was their ordinary diet. Though in general an appreciative person, he mentioned that “the inefficiency of the catering industry in Russia must be seen to be believed” (Report). They also must have been grateful for ordinary conversations in their second language as well as the infrastructure of services that they were used to: “Another surprise is the unpredictability of the outcome of any approach to a counter, window or door, at or behind which some service is expected or hoped for. I would have found it hard to remain fit and cheerful without the company and help of Mrs. Koopmans” (Report). But all in all, other than some tension in Moscow, they might have told their neighbors how friendly their reception was: “We met with great friendliness and cordiality from the professional colleagues visited, especially in Leningrad and in Novosibirsk.… We were also taken around by colleagues to see sites and museums in the most charming and relaxed manner” (Report). They were invited into a private house three times. All of them, Koopmans added, were the houses of mathematicians.
However, compared with his enthusiasm before his trip, there was also a sense of disappointment. He was not very impressed by many of the economists, specifically the senior ones:
My discussions were most free and spontaneous with mathematicians. I also had quite informal and frank discussions with the younger mathematical economists. The exchanges with the (better informed) senior mathematical economists were relatively more affected by caution and hesitation. As a result, the flow of research ideas was better than that of information about their contemplated use in the Soviet economy. (Report)
In his letter to his colleagues in game theory, cited above, he spoke of “the hesitant and diplomatic arms-length attitude that affected many of my conversations with economists” (letter to Aumann et al.). But, most disappointingly, he did not receive any answer for his “pressing question” regarding the use of prices for economic planning. “I pressed and had persistently the question whether the mathematical methods were to be used to perfect centralized planning.… The vagueness of the answers received suggests that this is still under discussion, if not dispute” (Report). Thus, Koopmans came home with mixed feelings—appreciation for some economists, but also the assurance of the superiority of US economists. His reckoning of the quality of Soviet economists was cool: “At the Moscow CEMI I met about 25 professional people in discussion, in Leningrad about 15, in Novosibirsk about 20. In each of these places I was impressed by the professional competence of about 8 or 10” (Report). Regarding the question of whether or not one should continue these sorts of meetings, he was cautious:
It is recommended that exchanges in the field of mathematical economics (in the wide Soviet sense of the term) be continued. It might be good if the next person to go were more interested in algorithms and in industrial applications of mathematical programming than I have been, because of the strong Soviet interest in these problems, and because they can be discussed by our Russian colleagues with less inhibition. (Report)
The major problem, according to Koopmans, was related to the organization of Soviet science; it lacked a peer process, which was the central dogma of the US postwar organization of science:
The main problem seems to be the absence of clear standards of quality of work in appointments and in publication. (It is somewhat like the early days of the Econometric Society, when one encountered a mixture of talents, with Ragnar Frisch, Charles Roos, Irving Fisher, Louis Bean, Henry Schultz and Theodore Yntema, all marching in the same ranks). (Letter to Kaysen, February 8, 1977)
However, in all these criticisms, one group was excluded: the mathematicians. He added to the quality count of “8 to 10 competent scholars”: “These estimates do not include the ‘pure’ mathematicians I have met, who were universally of high caliber, and more favorable to the various applications of mathematics than their U.S. counterparts generally are” (Report). Here and there, the Cold War created the same social conditions under which mathematicians could live out greater intellectual creativity in economics than economists themselves. Thus, the hierarchy between mathematicians and economists, including their quarrels, of which Koopmans knew better than many other US scholars, applied in the very same sense in the Soviet Union: Mathematicians, whether he liked it or not, are the better economists.
His sense of disappointment is most apparent in his letter ten years after his visit to Carl Kaysen. There is little to learn from the Soviets, he remarks, there is only something to learn about them, specifically how they fail in efficiently applying mathematics to the planned economy:
The principal direct U.S. gain from further … exchange activities in economics that I perceive is not so much to learn from our Soviet colleagues (I have been disappointed on that score) as to learn about them, about their clumsy economy, and about the extent of their impact on management and planning of the economy, which likewise falls short of its potential. (Koopmans to Kaysen, February 8, 1977)
In the years to come, Koopmans became active in the booming field of comparative economic systems—an interest that gained shape only after and through his travel experience. Comparing the two grand systems within a framework beyond these systems was a lasting interest after his “pressing question” remained unanswered (Koopmans and Montias Reference Koopmans, Montias and Eckstein1971). Such blunt comparison, however, required more commitment to a shared epistemic sphere than he could expect from his Soviet colleagues, as Koopmans must have learned from his travel.
In 1970, Koopmans traveled again to the Soviet Union for a conference organized by Kantorovich. He revisited some of the same places and people, in addition to the Institute of Automation and Telemechanics, the research institute of GOSPLAN, and the Laboratory of the Institute for the Management of the National Economy that was about to be taken over by Kantorovich. The field of mathematical economics had grown enormously in the previous five years and was no longer limited to CEMI. Even the Institute for World Economics and International Relations of the Academy of Sciences, a former Marxist bastion, launched a department for econometrics (headed by Stanislav M. Menshikov). At the conference, among the papers presented by his Soviet colleagues, he reported, were papers of equal quality to the best work in Western mathematical economics.
This overlap in theoretical focus in the mathematical economics of “East” and “West” may well leave the authorities in “East” a little impatient. However, I regard this particular development as important and beneficial for two reasons: (1) Good abstract “pre-institutional” economic theory ultimately promotes good economic policies in any system, and (2) meanwhile a climate favorable to exchanges of ideas and visits and to collaborative efforts is being established. (TKP, carton 22, “Report to the National Academy of Sciences on an Exchange Visit to the USSR, June 10–July 4, 1970”)
Koopmans, despite his many disappointing travel experiences, held on to his initial vision of pre-institutional economic theory that could bring about the convergence of the two intellectual spheres across the Iron Curtain: Soviet and Western economics might free themselves from their political bonds and merge into one unified effort of searching for the same truth in the same terms.
It is recommended … that exchanges in regard to mathematical methods in economics be pursued further and intensified because of a mutually perceived mutual interest in problems in this area. If difficulties arise or have arisen in regard to exchanges in other areas of economics (including that of the comparative evaluation of economic systems) the thought is expressed here that it may be best not to let exchanges in an area where they are desired by both sides suffer from difficulties encountered in other areas. The most important reason behind this thought is that the fundamental unity of all economic thinking will in time bring about the spread of ideas from areas in which exchanges are at present favored into other areas. (ibid., emphasis added)
Free-floating economic knowledge was in Koopmans’s mind as aloof were the hopes that he projected into it while ignoring his personal experiences regarding the asymmetry of the two regimes of economics in the East and West. In 1975, Koopmans and Kantorovich received the Nobel Prize, and thus Koopmans’s vision was publically confirmed: Two economists from the East and West joined in the “fundamental unity of all economic thinking”—though one might have been a little more at ease with the situation than the other.
