Wassily Leontief Applications

Wassily Leontief Applications

A proper account even of Leontief's own refinements and applications of input—output would require more space than we have available. Leontief and his colleagues initially calculated 44-sector input-output tables of the American economy in 1919 and 1929. For analytical work the sectors were aggregated to ten, and even then the calculations, made with punch-card equipment, were laborious. The development of computer technology after the Second World War eliminated manual calculations. Much larger models were computed within hours or minutes, and practical applications multiplied.


Leontief had by this time developed ways of incorporating the foreign sector into his model. The prices side of the model had also been developed and could be used to show, for example, the effects on prices throughout the economy of changes in wage rates or rates of profit. A dynamic version of the input-output model incorporated the investment required to augment the capital stock and allow the economy to grow (see Appendix), and inter­regional models were introduced. In 1948 Leontief became the founding director of the Harvard Economic Research Project, which was devoted to furthering input-output analysis. Several governments embarked at this time on the compilation of input-output tables, including the American government, whose first table, for the year 1947, was published in 1949. Input-output techniques came to be widely used in the preparation of plans for developing countries. The model did not however find universal acceptance. In the USSR and Eastern Europe it was regarded with suspicion and hostility as an economic tool which ignored the creativity of the socialist economic system, and no input-output tables were prepared in the USSR until 1958. Ironically, ideological obstacles also hampered government work on input-output in the USA. During the Eisenhower presidency, input-output was viewed as the thin end of the wedge of economic planning, and government-supported work was stopped.

Leontief himself demonstrated the versatility of input-output techniques with his influential work in the 1950s on the factor content of American foreign trade. Standard trade theory implied that the United States-a capital-rich country-would export capital-intensive and import labour-intensive commodities. Leontief, using his input-output model, demonstrated the reverse, that American export industries were less capital-intensive than American import-competing industries. This result, which became known as the Leontief paradox, was a watershed in international economics which called forth a number of desparate attempts to salvage the traditional theory. Leontief himself was highly critical of some of these efforts, and his own explanation of the phenomenon, which was in terms of the higher productivity of American labour, found a later echo in terms of 'technological' theories of trade flows and theories explicitly incorporating human capital in the analysis.

Leontief's most recent and most ambitious input-output model is one covering the whole world economy prepared for the United Nations. Work on this model started in 1973, and the Original focus was a projection of the impact of economic activity on the environment. The input-output model can be adapted in a conceptually simple way to incorporate this aspect. Each sector produces not only its normal output but also a number of pollutants, in quantities which are directly related to the scale of activity in the sector, (the assumption of proportionality or of fixed coefficients seems particularly appropriate here). The model is completed by incorporating a sector or sectors for the abatement of pollution, which has associated inputs and which destroys (produces negative quantities of) pollutants. Thus amended the model can immediately identify the total output of all sectors required to produce a specified bill of final goods while limiting pollution to a specified level.

However the scope of the United Nations study was extended to an investigation of the future of the world economy. The world was divided into 15 regions and an input-output model was developed for each region comprising 45 sectors, of which 22 are manufacturing; eight pollutants are also distinguished and five of the sectors are pollution-abatement activities. Within each region current and capital input coefficients were projected for the years 1980,1990 and 2000; from these the total input requirements (and pollution outputs) can be calculated for any rate of growth and structure of national income. The regions were also linked by trade and capital flows expressed in money terms; the calculations thus required the forecasting of relative prices, again using the input-output model to project the changes in costs associated with changes in factor prices and technological change.

The model as set out above cannot define a trajectory for the economy. To do so additional assumptions are needed, particu­larly on rates of growth of population and of output in each region. The study proceeded by developing scenarios based on alternative combinations of growth rates for the regions. The results are scarcely encouraging for those concerned about the income gap between rich and poor countries. If growth rates of developed countries are simply extrapolated, and if the assump­tion is made that developing nations achieve the growth-rate targets adopted by the United Nations for the Second Develop­ment Decade, then the income gap between rich and poor countries would remain in the year 2000 what it had been in 1970-about 12 to 1. Yet the United Nations targets for the 1970s proved to be hopelessly unrealistic, particularly for countries poor in natural resources, so the assumptions even of this scenario are — as Leontief acknowledges - unrealistic.

The second major scenario models the implementation of the New Economic World Order, based on a greater willingness of developed nations to open their markets to exports from develop­ing countries and on more generous capital flows from the former to the latter. The effect of these policies would be to reduce the income gap to 7 to 1 by the year 2000. Leontief concludes that 'the principal limits to sustained economic growth and economic developments are political, social and institutional in character rather than physical' (Leontief, 1977b, pp. 10-11). He is however acutely aware of just how serious these obstacles are.

Leontief continues to work on the World model, although lack of financial support has prevented him from developing it as he would wish. In the past two years he has used it to investigate the impact of cuts in military spending, in a study financed by the United Nations. He has also developed the treatment of popu­lation variables in the model (Leontief, 1979), and other studies are in progress. Work on adaptations of the model is going on in other countries as well, including the USSR.

As this brief account indicates, input-output analysis has become a powerful and widely-used tool for applied economic research. An input-output model forms the core of forecasting models of national economies, such as the Wharton School econometric model of the United States or the Cambridge growth project in the UK. Centrally-planned economies have now abandoned their early hostility and use input—output models extensively, particularly for longer-term planning. Business firms also make much use of the model. As Leontief wrote in 1976, 'for a fee one can now purchase a single row of a table showing the deliveries of a particular product, say coated laminated fabrics or farming machine tools, not only to different industries but within each industry to individual plants segregated by zip code areas' (1977a, p. 152).

Within the economics profession, however, input-output has been less influential. Certainly, linear models of the economy abound. Many authors employ in their theoretical work so-called generalised Leontief techniques, or representations of the economy which allow each sector a choice from among a number of discrete techniques, each with different fixed input coefficients (see for example Dorfman et al., 1958). Sraffa (1960) has in­vestigated theoretical properties of linear systems. What these works lack, however, is the potential for implementation in the real world to solve practical problems. The increased generality of many such models is purchased at the cost of making them inapplicable in practical work.

It is his emphasis on empirical implementation which marks Leontief out amongst contemporary economists. In 1970 Leontief chose the occasion of his presidential address to the American Economic Association to deliver an unfavourable assessment of recent trends in economics. In particular he criticised economists for establishing a scale of professional values which gives high marks to theoretical (usually mathematical) contributions and low marks to those engaged in the painstaking work of develop­ing the data base from which well-founded practical conclusions can be drawn. As a result, the means with which economists try to solve their problems are 'palpably inadequate', and economists display a 'continued preoccupation with imaginary, hypothetical rather than observable reality'. This scale of values is passed down from generation to generation by bias in the recruitment of academic personnel. Leontief calls instead for serious efforts to collect more accurate and more recent data and for a willingness to cross frontiers and collaborate with other disciplines, such as engineering and sociology. He believes, moreover, that much of what economists are trying to do is done much better by operations researchers or management scientists.

Views similar to Leontief's were expressed at the same time by several leading economists working in different areas of the subject. Nearly ten years later, however, in 1979, Leontief had seen little improvement in the situation. Academic journals are still replete with articles using complex mathematics to prove trivial results on unrealistic assumptions. Leontief now believes that it would serve little purpose for him to express himself again on the subject and that the major contribution he can make towards furthering his own conception of economics is the continuing example of his own work.

Before we consider in detail Leontief's views on more explicitly political questions, and particularly his controversial espousal of national economic planning in the United States, it may be worth outlining his other contributions to economics. Two topics here deserve particular attention-Leontief's work on aggregation and his perceptive and often critical remarks on Keynesianism.