A une époque où la majeure partie de la recherche économique consistait en « expériences de pensée » conduites de manière isolée, les recherches en input-output de Leontief faisaient appel à un travail d’équipe afin de rassembler des données et de réaliser des calculs. Il est d’ailleurs vraisemblable que les calculs input-output furent les premières utilisations des ordinateurs en sciences sociales. Ainsi, à la différence des autres branches de la discipline, l’analyse input-output était particulièrement « intensive en capital ». De plus, en raison des changements technologiques rapides, le capital matériel de recherche en input-output était soumis à une obsolescence très rapide.
Ce court essai fait la chronique de l’accumulation et des déplacements géographiques de l’équipement matériel qui a accompagné les recherches de Leontief depuis la fin des années 1940 jusqu’au début des années 1970. Le déménagement du matériel de Brandeis jusqu’au Japon eu lieu plusieurs années après, bien après que Leontief ait transféré ses activités à l’université de New-York.
Mots-clés :analyse input-output, capital de recherche scientifique, obsolescence, équipements de la recherche, Leontief (Wassily)
Keywords :input-output analysis, scientific research capital, obsolescence, research hardware, Leontief (Wassily)
1Tens of thousands of years hence archaeologists, sifting through the ancient artifacts at a remote site in what was once Japan, may be tempted to rewrite the history of economic thought in the twentieth century. While most of the artifacts are beyond recognition, these seem to be the remains of an early input-output project. Could it be that the Russian-born (or German-born?) American economist Wassily Leontief, the originator of input-output analysis, conducted his seminal research here on this island off the east coast of Asia? How else might one explain the presence of files, punched cards, tapes and assorted furniture in this miraculously preserved underground space?
2Actually Leontief did conduct his seminal research in the United States. While he had cordial relations with many talented colleagues in Japan his visits there were brief and primarily ceremonial. The artifacts of his 20th century activities were sent to Japan because he agreed to donate them to a Japanese input-output society in exchange for their financial support for his research activities at NYU.
- 1 This account covers only a fragment of U.S. research in input-output in the 20th century. It omits (...)
3Why dwell on “stuff” when ideas are what really matter? While human talent and ideas are at the heart of economic inquiry, physical capital also plays an essential role in modern research. Every scholar needs a place to think and to write and proximity to others enables collaboration and cross-fertilization of ideas. Equipment matters as well: in recent times computers have played an increasing role, providing access to information and enabling complex computations. Thus a review of the loci and the hardware of research is far from irrelevant in accounting for its course. While human capital remains far more important than physical capital in the research process, neither is effective without the other. This essay describes the changing character of space and equipment dedicated to input-output research in the early days, i.e., between 1947 and 1970 at Harvard and then at Brandeis from the early 70’s through the late 1980’s1.
4Last summer 1583 Massachusetts Avenue, a large grey wooden house just north of the Cambridge Common was moved 3 blocks north of its original location to make room for an expansion of the Harvard Law School. Moving a three-storey wooden structure with a footprint of roughly 1,200 square feet three blocks down a major thoroughfare took weeks of preparation and disruption of traffic. It would have been far simpler and cheaper to dismantle and replace the building but the Cambridge Historical Commission ruled that it must be preserved. It would be interesting to know what history the Commission had in mind: the building was erected in the late 19th century and I found out only recently that it built as a private home by a friend’s grandfather. Chances are good that it was preserved as a fine example of late 19th century architecture, although its beauty was concealed under a drab coat of grey paint for the years when I knew it best. Now it has been restored to the multicolored splendor of its period of origin but remains unoccupied. There is no plaque to indicate its historical significance for economists.
Figure 1: 1583 Massachusetts Avenue, offices of the Harvard Economic Research Project from 1955 to 1972.
Input-output materials were relocated from Harvard to Brandeis in 1972. These were augmented by data and other research paraphernalia used in building the United Nations World Model.
Picture by Sarah Carter Clunan.
5This building was, for close to 20 years, the home of the Harvard Economic Research Project (HERP), directed by Wassily Leontief and devoted to research in and applications of input-output analysis. For economists familiar with Leontief’s work the building represents the golden age of input-output analysis and the research infrastructure associated with it.
6The Project was started in 1947, a decade before the move to 1583, by Leontief, the inventor-originator of input-output analysis. In contrast to most economic research, input-output analysis called for detailed data on interindustry transactions and the solution of large sets – in those days perhaps close to 100 – of simultaneous equations. Because the work involved enormous amounts of quantitative data and computations that were relatively ambitious, it represented a new departure in economic research. In the late forties, the time of HERP’s birth, the standard mode of economic research featured a solitary economist developing broad principles that might be explained using a few illustrative numerical examples. A comfortable office equipped with pencil and paper, an accessible library and a typist to process manuscripts provided adequate infrastructure for mainstream economic research. At the time that he was invited to join the Harvard faculty, Leontief anticipated that his work would require more than the efforts of a solitary thinker and asked for a modest grant of $1,500 to pay for research assistance for gathering data and computing. The senior faculty was puzzled and skeptical but the request was granted.
7By the time Leontief retired from Harvard in 1974, it had become clear that input output-analysis, or any variant of multisectoral general equilibrium analysis, was of necessity a team effort. Gathering the data and constructing a table took years of work on the part of government agencies. While the construction of input-output tables is now a standard component of national accounting it took many years of collaboration involving the Census Bureau, government economists at the Bureau of Labor Statistics, the Bureau of Economic Analysis and Leontief and his associates at Harvard to establish meaningful and consistent conventions for an input-output table. Should transactions be measured in producers’ or purchasers’ prices? What about transportation costs? How should inconsistencies in reporting best be treated? Which conventions would best serve the research community and also the policy makers who might take advantage of this information.
8It’s not clear whether Leontief himself realized, as he launched his new Project, how radically this new mode of research would differ from the single-investigator standard. The Project grew incrementally from a one-man show to a 3-ring circus. Initially Leontief depended heavily on a single secretary-assistant, Bernadette Drolette, who, working in the secretarial office of his professorial suite, typed his letters and manuscripts and also assisted with computations, wielding a large, conspicuous slide rule. Slide rules were standard equipment for quantitative research in most fields in the early days. We all relied on them, sometimes supplemented by Friden or Marchant electric calculators, noisy machines that cost much more, in inflation-adjusted dollars, than today’s much more powerful, versatile and quiet personal computers.
9In the late 1940s Leontief was awarded a grant by the Ford Foundation to continue his input-output studies. Harvard assigned a few extra offices on the third floor of Littauer, surrounding Leontief’s faculty office, for “The Project”: for research assistants and a growing number of junior and visiting faculty members interested in this new line of research.
10However, Littauer became prime academic real estate as the Harvard faculty grew in the postwar student bulge. “Extra” offices were no longer available and it soon became clear that Leontief’s collaborators and support staff would have to be satisfied with some “subprime” offices in obscure niches of the University. I remember being assigned to an office in the basement of the Vanserg building, about 10 blocks East of Littauer in back of the biology laboratories. It was a “temporary” building, erected as such during World War I. In the late forties it housed (in addition to our offices), the Mathematics library and a cafeteria, the latter just above my office. The rolling of carts across the cafeteria kitchen was noisy and vibrations rattled the whole building. One day I returned from the library to find that the fluorescent ceiling light fixture had fallen onto my desk. The room was littered with shards of broken glass from shattered fluorescent bulbs. I was pronounced “lucky” and assigned to an office in another building. (As the centennial for that “temporary” World War I building approaches, it is still in use, now as Harvard’s piano repair shop.)
11My new office was on the top floor of Harvard’s Semitic Museum where junior staff shared a few otherwise idle rooms. As I recall, it was a four storey building housing impressive ancient and massive stone sculptures. Outside of a very few delegations of visiting school-age children accompanied by nuns, we were the only live humans in the building. It was only a five minute walk to Littauer but we had very little interaction with Leontief and there were no accommodations for visiting scholars during that period.
12In the early 1950’s Harvard gave us a suite of offices on the top floor of Dudley Hall, a former dormitory designed to accommodate affluent students and their valets in the late 19th century. It was an elegant building with balconies overlooking a central indoor courtyard. Other floors of the building housed the Russian Research Center and Linguistics. Each suite had two spacious offices and a large bathroom with original plumbing fixtures including a deep 8-foot long bathtub. Leontief retained his Littauer office but occupied a second “research” office in the Dudley Hall complex and it was here that the “Leontief Paradox” computations were conceived and prepared. Executing the computations required many trips to Harvard’s 7094 mainframe at the computer center, located about 6 blocks away, near the Law School.
13The Dudley Hall suite was luxurious compared with earlier space allocations but was poorly configured for research offices and still far from the computer center. Serendipitously, in 1957 Harvard decided to demolish the building to make way for Holyoke Center, a major office complex that now houses administrative offices, the health service, restaurants, a bank and stores that serve the university and the community. Apparently the time was right for our project: government and foundation funding began to flow in and Harvard agreed to renovate a roomy old wooden building at 1583 Massachusetts Avenue for our use. The building was at the edge of the Harvard Law School complex and less than a block away from the computer center.
14It was a large “double” house, with twin front entrances, built for family occupation by the grandfather of a contemporary of mine, probably in the last decades of the 19th century. For most of this period its ample offices were fully occupied by Leontief (who also maintained an office in Littauer, where the Economics Department was officially housed), his research staff, visitors from other countries, and a sizeable contingent of administrative and secretarial staff, computer programmers and research assistants.
15One room was set aside for electric calculating machines and for card punching equipment. The Harvard mainframe computer, first an IBM 7094 then upgraded to an IBM 360 was housed nearby, about a block and a half away and researchers and programmers carried their boxes of cards back and forth for submissions, corrections and resubmissions. The basement housed a library of input-output and related books and papers along with file drawers full of documentation. At the back of the first floor there was also a full kitchen and a conference room with a long table shaped like the keel of a boat and a full wall of blackboard. 10:30 coffee hour was the occasion for presentations by staff and/or visitors on some days, and for informal exchange and gossip on others.
16This substantial infrastructure supported Leontief’s work and that of research staff. At any given time there might be six or seven full time research associates, a similar number of full - and part-time research assistants, a librarian, a professional computer programmer and a secretarial staff, in addition to faculty members and graduate students working on related projects. Each year the Project hosted distinguished visitors from many different countries who contributed very significantly to our shared agenda. The unpublished progress report for the academic year 1958-59 lists 56 international visitors from 16 different countries. Some stayed for only a few days, others for a week or two and some for much longer: up to two years. Together they could exploit an infrastructure that included a specialized library, organized data, access to the latest computing equipment, a support staff experienced in input-output data and computations and easy access to experienced colleagues with similar interests. It is hard to overestimate the advantages of working near to other scholars and research assistants who knew the data sources and the jargon, secretarial staff equipped with expertise in typing mathematical expressions and equipped with the hardware it required and opportunities for interchange with colleagues with similar interests. Coffee hour offered a natural opportunity to float new ideas and solicit feedback and a stream of domestic and foreign visitors contributed fresh ideas, questions and perspectives.
17Certainly HERP was not the first example of a centralized research organization in economics. The National Bureau of Economics Research, while less specialized, offered, and continues to offer, similar advantages and the same can be said of the Cowles Commission, the Rand Corporation and many other more and less specialized research organizations. The current objective is not to claim primacy for HERP as a specialized economic research organization but rather to focus on the phenomenon of research infrastructure and to illustrate how changes in technology and the evolution of scientific priorities can affect it.
18While, as noted above, HERP was far from a one-man operation Leontief remained the director and dominant leader of the organization throughout. In 1973 he passed the age of mandatory retirement at Harvard and prepared to depart for NYU where he accepted a post-retirement appointment. I had, by that time, accepted a professorship at Brandeis University in Waltham, Massachusetts, about half and hour’s drive from Harvard. This was before he had definite plans to start a new input-output project at NYU and so he asked me to take the HERP materials to Brandeis.
19Again, serendipitously, The Brandeis Economics Department had just moved to spacious quarters in the new Sachar Building and we had access to a very large storage room and a multi-tiered library space furnished with elegant Danish furniture. The HERP library was moved into the lavish space, then dubbed “The Airport Lounge.” Innumerable file cabinets containing input-output tables from all over the world, the detailed documentation of the 1939 and 1947 matrices, miscellaneous research reports and correspondence as well as dozens of boxes of computer cards representing all the input-output computations we had ever done at HERP, were loaded into the storage room along with miscellaneous machinery and furniture. Since our meticulous HERP librarian, Charlotte Taskier, had retired some years earlier without a successor, it arrived in unlabeled boxes to be unloaded and arranged by work-study undergraduates who hadn’t a clue... The storage room wasn’t quite as large as it first seemed!
Figure 2: Entrance to the Sachar International Center at Brandeis University.
Picture by Sarah Carter Clunan.
20One piece of equipment that didn’t come to Brandeis was a year-old, ponderous computing machine, about 5x4x1 feet, the latest technology that replaced our sluggish and noisy electric Marchants and Fridens. Karen Polenske, who was moving her work to MIT, reminded me that the bulk of the money to purchase this new machine came from her grant and that it was, therefore, rightfully hers. I chose not to dispute the claim. To my delight, I was able to purchase a 6x3x.5 inch (silent!) handheld machine for $75 the following year. Had I waited a few more years the handheld would have decreased to the size of a business card and been available free of charge to promote some event or other product. Sic transit…!
21As luck would have it, the Brandeis mainframe was a DEC PDP-10, whose input mode was tapes, not cards, and, while all the boxes of cards were later transported to Japan, it’s unlikely that they were ever read again. As we should have anticipated, the comfortable and stylish chairs and other small items of furniture in the Airport Lounge were spirited away, presumably by students interested in increasing the comfort of their on – or off – campus living quarters. Similarly, many of the more popular books and journals from the HERP library, placed on the open shelves without protection of a librarian, also disappeared. Perhaps they were to be read by the thieves in the very chairs that vanished in the same period. We’ll never know!
22Before there was time to put the stored materials into proper order our attention was diverted to another major challenge: Leontief was invited by United Nations to do an input-output projection of the world economy to the end of the 20th century. He presented the general idea of the world model as his Nobel lecture in 1974, the year he left Harvard for NYU. HERP had been dissolved and he had not yet organized a research team at NYU. Thus he called on me to assume responsibility for building and implementing a full-scale world model at Brandeis. (Vide Rumpelstiltskin!) Fortunately, Peter Petri agreed to come to Brandeis to share responsibility for the project. Without Petri there would have been no United Nations World Model. Since Brandeis had no graduate program in economics until some 20 years later, we mounted this major research undertaking using a professional computer programmer, a secretary and part time undergraduate research assistants. The enormous challenge of this project reduced the priority of organizing and archiving the HERP library and its artifacts, still stashed away in the storage room. The new project called for fresh data and the current computer technology had substituted noisy dumb terminals and tapes for punched cards. When Leontief requested the transfer, we simply boxed the contents of the storage room and shipped it to an agent in Japan. We were never informed of its final destination.
23Of course, the useful life of the tapes and dumb terminals was also short. Except for a few supercomputers, campus mainframes no longer exist. They were superseded first by desktop computers which will presumably be displaced before long by smaller and more powerful portable computing devices. Data are posted and transported on the Web, and boxes of cards and tapes, along with the devices that read them, have become quaint exhibits in computer museums. It seems that archaeologists need not wait tens of thousands of years to begin their research. The “Leontief Library” in Japan has already become an ancient artifact.
1 This account covers only a fragment of U.S. research in input-output in the 20th century. It omits the essential and sizeable investment by government agencies in gathering the basic and essential information about industry production, interindustry sales, purchases, employment and investment embodied in input-output tables. Nor does it report on the input-output related research of scholars in the U.S. and throughout the world, including that of Leontief and his associates at New York University, although that work was substantial and significant.Haut de page
Table des illustrations
|Titre||Figure 1: 1583 Massachusetts Avenue, offices of the Harvard Economic Research Project from 1955 to 1972.|
|Légende||Input-output materials were relocated from Harvard to Brandeis in 1972. These were augmented by data and other research paraphernalia used in building the United Nations World Model.|
|Crédits||Picture by Sarah Carter Clunan.|
|Titre||Figure 2: Entrance to the Sachar International Center at Brandeis University.|
|Crédits||Picture by Sarah Carter Clunan.|
Pour citer cet article
Anne P. Carter, « Leontief’s “Stuff” », Œconomia, 1-1 | 2011, 49-58.
Anne P. Carter, « Leontief’s “Stuff” », Œconomia [En ligne], 1-1 | 2011, mis en ligne le 01 janvier 2011, consulté le 25 avril 2017. URL : http://oeconomia.revues.org/1823 ; DOI : 10.4000/oeconomia.1823Haut de page
Haut de page
Les contenus d’Œconomia sont mis à disposition selon les termes de la Licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.