The Web Book of Regional Science  

5 International Developments

The first government agency to undertake the construction of a full-scale national input-output table was the Bureau of Labor Statistics of the Department of Labor. This effort resulted in the publication of a 50-sector table of interindustry relations in the United States and of a much more detailed 200-sector table with finer industrial and sectoral classifications.1  The hypothetical table described in Chapter 2 (Table 2-1) is modeled after the 1947 national table published by the Bureau of Labor Statistics. To construct this table, a separate Division of Interindustry Economics had been established in the Bureau of Labor Statistics. An important result of this early work in input-output analysis was a projection of the U.S. economy to 1950.2

The work of the Division of Interindustry Economics attracted widespread attention among economists and businessmen. Unfortunately, in some quarters it was considered "controversial." Some businessmen were said to have viewed the program as a step toward "push button planning" and a threat to private enterprise.3 Appropriations to the Department of Labor were curtailed and, while the Department of Defense had sufficient funds to continue work on input-output analysis, the decision was made by a Deputy Secretary of Defense to terminate support of input-output studies after November 1953.4 There was no further work on interindustry analysis by the United States government until after the Census of Manufactures of 1958. At that time, the Office of Business Economics of the U.S. Department of Commerce undertook the construction of a new input-output table for 1958 which was published toward the end of 1964.

Research on input-output analysis continued at the Harvard Economic Research Project and at other universities, largely financed by foundation funds. But the construction of a national input-output table is a major statistical effort. While private research organizations are admirably suited to conduct research on input-output analysis, and in many cases to conduct regional input-output studies, the statistical and financial resources of government agencies appear to be a prerequisite for the successful construction of national tables. And because of the curtailment of funds in 1953, there was a period of more than five years during which government agencies in the United States could not engage in such analysis.

Although empirical work on input-output analysis languished in the United States, it surged ahead in other countries. And the rapid spread of input-output analysis throughout the world stimulated a large number of theoretical studies to complement the empirical work being done. By 1961, a partial bibliography of input-output studies—both empirical and theoretical —published by the United Nations ran to 222 pages, and agencies in about 40 countries were involved in interindustry studies.

As early as 1951 there was sufficient interest in this new analytic technique to stimulate an international conference on inter-industrial relations. This conference, which met in Driebergen, Holland, brought together economists interested in the theoretical, the statistical, and the computational problems of interindustry analysis.5 A second conference was held between June 27 and July 10, 1954, at Varenna, Italy.6 A third international conference was held in September 1961 in Geneva. Economists and statisticians from more than 41 different countries participated in this conference. For the first time, representatives of the Soviet Union and of other socialist countries, as well as planners from underdeveloped countries, participated in the international input-output conference.7

The first international conference dealt largely with the empirical implementation of input-output systems. The major emphasis of the second conference was on statistical and computational procedures and problems. The central theme of the third conference was the application of input-output analysis to projection and developmental planning.8 Thus, during the decade spanning the three international conferences, there was a marked shift from emphasis on the problems of constructing input-output systems to the application of these systems to a variety of economic problems.

Input-Output Analysis in Planned and Unplanned Economies

The first empirical application of input-output analysis was in the United States, an unplanned economy which depends upon market forces for the allocation of resources and the distribution of income. The input-output system is not a tool developed by "planners" with the intent of substituting another form of economic organization for the market system. Indeed, the early work in interindustry analysis was oriented toward a market economy. The objective was to measure, as precisely as possible, the impact upon the economy of autonomous changes in final demand. Within the framework of a free-market economy, the input-output analyst is not particularly concerned about the causes of changes in final demand. These are "given." And once they have been estimated, the input-output system will show the levels of activity which will have to be met within the endogenous sectors to sustain this level of final demand. The input-output system as such is not a planning tool—it is an analytical tool. But while it was developed within the framework of a market economy, it soon became apparent that this tool could be applied to other types of economy systems as well.

Input-output analysis in partially planned economies. Before World War II, there was considerable debate among economists about the virtues of "planned" versus "unplanned" economies. Much of this debate was conducted in polar terms. One either talked about a "planned" economy, by which one meant a totally planned economy (of which the Soviet Union was generally considered the prototype), or one talked about an "unplanned" economy by which was usually meant a laissez-faire system in which all economic decisions were made by the invisible hand of the market place.

 Experience since the end of World War II has shown that much of this debate was of purely academic interest. Like so many controversies which pose two absolute conditions as mutually exclusive alternatives, this one was shown to be of relatively little relevance to the real world. During the Second World War, there was a great deal of "planning" in all of the countries which participated in the hostilities. And after the war, many of the countries in Western Europe continued to engage in what might be called "partial" planning. This type of planning may or may not be associated with some degree of socialism—some degree of government ownership and operation of the major means of production (usually basic "heavy" industries). In Great Britain, for example, the government nationalized some industries; in other cases, such as France and Italy, there was little or no experimentation with socialism, but in these and other countries there were, and continue to be, various experiments with different kinds of "planning."

 The notion of "indicative planning" as it is practiced in France was introduced in Chapter 3 as part of the discussion of consistent forecasting. France is one of the countries whose government has engaged in a certain amount of economic planning in recent years within the framework of a private-enterprise economy. One of the analytical tools which has been prominent in French indicative (or noncoercive) planning is an input-output model which is geared to the French system of national accounts. Long-term projections (five to ten years) of final demand are made, assuming several different rates of growth. These projections provide a set of production targets which will have to be met if the final demand figures are to be realized.

 The General Commission on the Plan has the basic responsibility for the initial forecast. Other government agencies also participate. The initial plan is then submitted to committees (e.g. the Committee on General Manufacturing Industries) made up of private businessmen. The committees, which include a few civil servants and trade unionists but are predominantly made up of businessmen, review the plan.

 In general, the committees examine the sectoral projections, which are quite aggregated in the French input-output system. They then attempt to determine the detailed industrial outputs which will be required to meet these aggregated projections. At this stage there is still a considerable amount of flexibility in the procedure. Reports are submitted to the General Commissioner for the French plan, and on the basis of these reports the Commissioner may alter the initial projections before the Plan becomes "official." Before this is done, however, the Plan is reviewed by the Economic and Social Council; it is then sent to Parliament. Following this, the official Plan is published, but it "does not imply any obligation, nor any sanction."9

 There is, of course, more to indicative planning in France than this sketchy description suggests. Demographic trends, fiscal policy, the international balance of trade and payments, and other factors are taken into consideration in the preparation of the Plan. The input-output technique, however, is the central analytical tool in French indicative planning. It is also important to stress that in making the long-term projections there is close cooperation between government agencies and committees made up predominantly of private businessmen. The businessmen have recognized that the reduction of uncertainty contributes to the stability of their operations. The effectiveness of this joint action is demonstrated by the rapid growth of the French economy in recent years, and the extent to which "full" employment has been maintained. As is true of most of the industrial nations of Western Europe, France has had a remarkably low unemployment rate in recent years. This is not a suggestion that indicative planning is a panacea for all of the ills of industrial society. The only point to be made here is that the input-output system has proved an extremely useful analytical tool in a partially planned economy. It should be noted in conclusion that European nations which have engaged in relatively little "planning" (such as Italy) as well as others which have engaged in more "planning" than France (such as Great Britain) have made extensive use of input-output analysis.10

 Input-output analysis in a completely planned economy. There is no such thing as a totally planned economy in which every transaction is projected in advance. There is, however, central planning; the leading practitioner of this method of allocating resources and distributing income is the Soviet Union. Much of the controversy referred to earlier in this chapter about "planned" versus "unplanned" economies centered on the question: Can economic planning work?

 The experience of the Soviet Union shows that central economic planning can work, although to an economist trained in the Western tradition the success of early Russian economic planning remains something of a mystery. The basic problem faced by policy-makers in Russia after the Revolution of 1917 was that there were no planning guidelines for the type of system they were trying to set up. The only economic theory, tolerated in the Soviet Union was Marxism, and as Leontief has pointed out, "Marxism, as an economic theory, is a theory of rampant private enterprise, not of the centrally guided economy."11 Nevertheless, under Stalin a series of five-year plans were promulgated, and whether or not the planned targets were achieved, the Russian economy entered an era of rapid growth. The basic principles guiding the early "planners" in the Soviet Union were simple. The objective was to produce as much as possible, consume as little as necessary, and use the surplus for investment to stimulate further economic growth. Investment decisions (how much to invest and where) were made by Gosplan, the central planning agency. It was this agency which decided investment priorities and production targets after consultation with directors of the major Soviet enterprises. All this was done without a basic analytical model and "so far as the Russian technique of economic planning is concerned, one can apply to it in paraphrase what was said about a talking horse: the remarkable thing about it is not what it says, but that it speaks at all."12

 Part of the problem faced by Soviet planners was that they were the victims of severe ideological constraints. During the Stalin era, the only economic theory they could use was that of Karl Marx. And while Marx had many penetrating insights about the operation of the capitalist economy, he did nothing to suggest how a centrally planned socialist economy would operate. More recently, however, some of the restraints upon Soviet economists have been lifted; it has been possible for them to study the analytical techniques developed in the West, and to begin to apply them to problems of Soviet economic planning. Until a few years ago, all references to "bourgeois economics" were highly critical. Leontief credits Oskar Lange, formerly a University of Chicago economist and now head of economic planning in Poland, with the introduction of a positive approach to econometrics in the Soviet Union and its satellites.13 To some extent, however, the new attitude toward analytical tools developed in the West is probably part of the "new freedom" which intellectuals in the Soviet Union have found in the post-Stalin era.

 There has been a great deal of interest among Russian economists in input-output analysis in recent years. American books and articles on interindustry analysis have been translated and widely circulated. As they have done in a number of other cases when they have "borrowed" ideas from the West, Russian apologists claimed priority in the invention of input-output analysis. This claim is based upon the publication of an article in a Russian economic journal written by Leontief while he was a student in Germany.14 A more legitimate claim to a related technique was found in the earlier-neglected, pioneering work of L. V. Kantorovich.15 The early paper by Leontief and Kantorovich's work established the intellectual respectability of input-output analysis in the Soviet Union. While empirical work on input-output analysis in Russia lags behind that in a number of other countries, it is apparent that rapid strides are being made.  

A great deal of secrecy surrounds the work on input-output analysis in the Soviet Union, but enough information has leaked out to permit the reconstruction of a 38-sector interindustry  table for that country based on 1959 transactions.16 A number of scholars have analyzed Soviet interindustry relations on the basis of partial data released by the Russian government. Recently, for example, Herbert S. Levine has contrasted the ways in which input-output analysis is used in centrally planned and in free-market economies.17  Somewhat more detailed information about Soviet input-output analysis (as well as the reconstruction of the 1959 Soviet input-output table) are contained in a paper entitled "Economic Interrelations in the Soviet Union," published by the Joint Economic Committee of the U.S. Congress.18 These studies show that while Soviet input-output analysis is closely patterned after earlier work conducted in the free-market economies, there are some significant differences. These are principally differences in application rather than construction, and to highlight them it will be necessary to give a brief discussion of the Soviet planning process.19

 The construction of a short-term (annual) plan involves a series of stages. The initial stage consists of what Levine has called the "flow and counterflow of instructions and information." This is followed by establishment of the major objectives of the economy which are given to Gosplan by the nation's political leaders. Gosplan then details a series of preliminary production targets called "control figures." These targets are transmitted through the economic chain of command to the basic enterprises in the Soviet system. Having been given its production target, each enterprise then prepares a list of the material inputs it will need. The enterprise is not free to determine its inputs arbitrarily, but relates them to a set of "materials input norms" most of which are determined at higher levels of the economic chain of command.

 In drawing up the plan there are two sets of pressures at work. Those at the top of the planning hierarchy apply pressure to reduce input requirements. At the level of the enterprise, which is responsible for meeting a production quota, there are counter-pressures to add a little extra to actual input requirements.

 After the initial production targets and input requirements have been prepared, it is up to Gosplan to achieve an internal balance for the plan as a whole, to see that the output targets and the input requirements are consistent. To achieve this, Gosplan uses what has been called the "method of material balances." Essentially, this consists of setting up a series of accounts — similar to the balance sheet used by accountants in a free-market economy, but expressed in terms of materials rather than monetary units. On one side the sources of supply of materials are listed, and the other side lists the uses to which these materials are to be put. As Levine has noted, "it is only by the wildest chance that the two will be equal at the first balancing.”20 In general, the material demands will be greater than the available supplies. It is up to Gosplan to bring the demand and supply into balance, keeping in mind the production targets which have been given the highest priority by the nation's political leaders. Gosplan tries to work out the balance by a trial-and-error approach, or in more technical terms by following an iterative procedure.

 Even with this sketchy discussion it should be clear why input-output analysis has appealed to Soviet planning technicians. Given a detailed input-output table, Soviet planners could achieve an internal balance very quickly by using electronic computers. Input-output analysis can be conducted in "a language the computer can understand (something not accomplished by the material balances method)."21 As Levine has noted, much Soviet discussion of input-output analysis has emphasized the speed with which it can be accomplished. In a centrally planned system this is very important since it would permit the development of a series of plans from which the Soviet leaders could choose rather than the "often poorly balanced, late, single variant now constructed."22

 The major difference between the application of input-output analysis in a free-market economy and in a centrally planned economy can now be made explicit. In a free-market economy the input-output analysis generally starts with a set of final demands. Then, given an inverse matrix (the table of direct and indirect coefficients), total outputs can be computed for each sector. By relating these to the table of direct input coefficients, a new table of interindustry transactions for all processing sectors can be constructed.

 In a centrally planned economy the targets established are not final demands but total outputs. These include not only the final demands but also the interindustry transactions needed to achieve these final uses. It has also been suggested that in a centrally planned economy the input coefficients should not be empirically determined (as in a free-market economy) but that intersectoral balances should be based "on scientifically determined progressive input coefficients."23

 Up to now Soviet economic planners have used input-output tables based on empirically determined input coefficients, and for short-term planning they have assumed that most of these coefficients will remain stable. Treml has noted that "out of 4,260 input coefficients shown in the 1959 intersectoral balance only 500 were adjusted when the first planning balance was being prepared for 1962."24 It is evident, however, that some Soviet economists are thinking in terms of projected coefficients—a goal similar to that of input-output researchers in the Western world who are working on dynamic models.

 Input-output tables have also been constructed for a number of other planned economies. By 1964, most of the large countries in the Soviet bloc had prepared such tables. These include Yugoslavia, Poland, Hungary, East Germany, Bulgaria, Czechoslovakia, and Romania.25 In spite of this impressive array of input-output tables in planned economies, both the theoretical and empirical work on input-output analysis in these countries lags behind that of the Western world. Since economists in the planned economies started to experiment with input-output much later than their counterparts in free-market countries, their accomplishments "should be viewed against the background of ideological obstacles and resistance to change from many quarters."26

 A Value-Free Science of Economics

 Mathematicians and many kinds of physical scientists, such as physicists and chemists, work in areas which are essentially free of political ideology. They can use the same tools, and converse in the same language, whether they live in a society with a free-market or a. centrally planned economy. This has not been true in other disciplines, notably the social and biological sciences. It is not surprising, given the antithesis between communism and capitalism, that Soviet leaders would repudiate the entire body of economic theory from the Classical through the Keynesian schools. Indeed, as Leontief has noted, Soviet planners have operated without a theoretical framework since even Marxism is a theory of capitalism which contains no useful guides to the Soviet planner.

 The input-output model is independent of political, social, and economic systems. Unlike the models of the major schools of economic thought of the past it says nothing about how resources should be allocated and incomes distributed. It is a value-free system which can be applied in free-market, partially planned, or totally planned economies. An input-output analysis tells us nothing about what should be; it describes the economy as it is. Various assumptions can be made about changes in technical coefficients, in final demand, or in total gross output. Once these assumptions have been made, the system can be used to make projections regardless of whether resource allocation and in­come distribution are determined by market forces or executive decree.

 As noted, input-output analysis had to become ideologically acceptable before it could be used by economists in the Soviet Union. But this was due entirely to pressures exerted by political leaders; it had nothing to do with the objective reality of input-output analysis. Once the ideological barrier had been hurdled it was still necessary for Soviet economists to reconcile input-output analysis and Marxian theory. According to Marx, the total output of an economy consists of three parts: (1) the value of capital used up in a given period (which is considered to be "embodied labor value"), designated by the symbol c, (2) v which is the value of labor used in the production process, and (3) m which is "surplus value" or profit. The nation's total output therefore equals c + v + m.

 Although Western economists would not accept such a Procrustean classification of the factors of production, Soviet theorists have managed to "combine" the sectors of their input-output tables to conform with the Marxian classification.27 Most Western economists would consider the division of an input-output table into quadrants which show capital inputs and "surplus value" inputs respectively as an unnecessary ideological exercise. But this division has no influence upon the values of technical coefficients, final demand, and total gross output. It was no doubt necessary for Soviet theorists to do this to make input-output politically palatable, while Western economists have not been hampered by even such minor ideological restrictions. The important point is that input-output is a technique which can be applied to a wide range of economic problems independent of economic systems.

 Input-Output Analysis and Economic Development

 The versatility of the input-output model was emphasized by Tibor Barna in his introduction to the proceedings of the third international input-output conference: "In 1961, at the Geneva conference the attendance was some 240, with about 100 actively participating in the discussions. And they came from 41 countries; from capitalist and communist countries, from developed and underdeveloped countries alike. They represented an international fraternity of economists and statisticians, trying to talk a common language and trying to learn from each other irrespective of political divisions."28  Barna's remarks also indicate that the use of the input-output technique is not restricted to advanced, industrialized nations; it has become a major analytic tool in the important field of development economics. As such, the "underdeveloped" nations of the world have exhibited as much interest in input-output analysis as have those with highly complex industrialized economies.

 The decline of colonialism and the rise of a host of new independent nations after the end of World War II provided a powerful stimulus to the economics of development. Since they now control their own destinies, the people of these nations seek to raise their standards and planes of living. In this desire they have had assistance from some of the world's developed economies, notably the United States and more recently the U.S.S.R.

 The objective of economic development is to move in the direction of greater self-sufficiency. The word greater should be emphasized. Complete self-sufficiency would mean sacrificing the benefits of the international division of labor and exchange. But some of the world's underdeveloped nations have relied almost exclusively on imports for most manufactured products. If they are to achieve higher standards of living they must become more self-sufficient than they have been in the past. This implies industrialization, and the latter, in turn, requires imports of capital goods while an effort is made to create "import saving industries in other directions."29 if the latter goal is to be achieved there must be expansion of "structurally related" industries. In general, this means the development or expansion of industries which produce inputs for other import-saving economic activities.

 The process of economic development is not a simple one, and there are wide differences of opinion among economists about the historical causes of differential rates of economic growth among the world's industrialized nations.30 As a result of the pioneering work of Allen G. B. Fisher and Colin Clark, however, it is evident that the goal of higher real per capita income will be achieved only if there is a shift in employment from primary sectors (such as agriculture, forestry, and fishing) to secondary sectors (manufacturing, mining, and construction), and thence to the trades and services which are lumped together under the heading of "tertiary" activities.31 Thus economic development means the restructuring of an economy, and in an increasing number of underdeveloped countries it has been recognized that this process will be hastened if modern analytical tools are used to show how this restructuring is to be accomplished. It has been recognized that the input-output technique is ideally suited for analysis of the structure of development.32 As Leontief has noted, the "input-output table is not merely a device for displaying or storing information; it is above all an analytical tool."33 This is demonstrated by a number of papers in the Proceedings of the Third International Conference on Input-Output Techniques.34

 If a country wishes to industrialize it will try to adopt the structure of an advanced economy. It will try to move from a position of relatively weak interdependence to strong internal interdependence: "The process of development consists essentially in the installation and building of an approximation of the system embodied in the advanced economies of the U.S. and Western Europe and, more recently, of the U.S.S.R.—with due allowance for limitations imposed by the local mix of resources and the availability of technology to exploit them."35 Input-output analysis provides a map for this process of development. Comparison of the structural relationships in an underdeveloped economy with those of an industrialized economy will show the gaps that have to be filled. And an input-output table will show the effects—direct and indirect—of expanding a given sector or of adding new sectors to those already represented in the underdeveloped economy.

 Comparison of input-output tables for two or more economies (national or regional) is facilitated if the tables are standardized - if the rows and columns are arranged in a logical scheme rather than in the sequence prescribed by the official arrangement of statistics in various countries. This is accomplished by "triangularizing" the input-output tables to be compared.

 Triangularized Input-Output Tables

 The example of a triangularized input-output table to be discussed in this section is greatly simplified; it has been "precooked" in order to demonstrate certain principles.36 Triangularizing a real input-output table is a difficult task. It is not enough to arrange the rows and columns on the basis of zero and nonzero entries. The magnitude of the latter must also be taken into account, and the arrangement must follow some set of predetermined criteria.

 Table 5-1 is a hypothetical matrix of the processing sectors of an underdeveloped economy. The only difference between Table 5-1 and the processing sectors of Table 2-1 is that there were relatively few zero entries in the latter while almost half of the entries in Table 5-1 are zeros.


Table 5-1 illustrates "weak structural interdependence" as opposed to the strong interdependence illustrated by Table 2-1. We assume that this table was constructed on the basis of Standard Industrial Classifications followed by statisticians in the underdeveloped economy. The arrangement of sectors is based entirely on the customary way in which statistical data are ordered, and Table 5-1 shows no particular pattern of either dependence or independence among industries. It is however possible to rearrange the random distributions of this table into an order which has meaning. This has been done in Table 5-2. which is a triangularized version of Table 5-1. Two criteria were used in rearranging the hypothetical transactions table: (1) The sector with the largest number of zero output entries was placed at the top of the table, while each row below this has a smaller number of zero entries, and (2) the rows have been arranged so that final demand as a percentage of total gross output declines as one reads down the table. 

 An actual triangularized input-output table would not have the perfect symmetry of Table 5-2. The hypothetical table has been made symmetrical to illustrate the principles involved in standardizing input-output tables, and some of the reasons for doing so. Industry B is highly dependent upon other industries in the hypothetical underdeveloped economy for its inputs. But it sells nothing to other industries; its total output goes to final demand (including export sales). There are some intraindustry transactions, but these are the only sales within the processing sector.37

 At the other extreme, industry C buys nothing from other industries in the processing sector. More than half of its total gross output is sold to other processing sector industries, however, and only 43 per cent goes to final demand (including export sales). Thus industry B is an example of strong interdependence on the input side and industry C an example of strong interdependence on the output side. The table as a whole, however, shows relatively weak structural interdependence.38

 After the transactions table has been triangularized, technical coefficients, and direct and indirect requirements per dollar of final demand, are computed following the steps outlined in Chapter 2. The results of these computations for the hypothetical underdeveloped economy are given in Tables 5-3 and 5-4.



Table 5-4 differs in one important respect from its earlier counterpart, Table 2-3. The latter is a transposed inverse matrix while Table 5-4 has not been transposed. It is convenient when possible to transpose an inverted input-output matrix which then shows the production required from each industry at the top per dollar of deliveries to final demand by each industry at the left. It is not essential that this be done although it does make the inverted table easier to read. Table 5-4 was not transposed since this would have shifted the zeros below the diagonal and this might have been confusing. The notes directly below Table 2-3 and Table 5-4 show how each is to be read.


As noted in Chapter 3, the inverted Leontief matrix, or table of direct and indirect requirements per dollar of final demand, can be used to forecast the total impact on an economy of changes in final demand in one or more sectors. In the following example we will show how such a table can be used for development purposes. To do this we will assume the following changes in final demand for each of the industries in the hypothetical underdeveloped economy.

Assumed Changes in Final Demand


Final Demand
Final Demand
Per Cent


Given these changes in final demand we can (as in Chapter 3) project all changes which will take place within the processing sectors of the table. Such projections for the triangularized matrix are given in Table 5-5.


We have assumed that the final demand for industry B's products goes up 20 per cent, or in absolute terms (expressed in U.S. currency) four billion dollars. What effect does this have upon production in the other industries represented in the hypothetical transactions table? Reading down column B and adding the differences between the original and projected transactions, we find that total sales to industry B (including intraindustry transactions) go up 3.6 billion dollars. The increase in transactions within the processing sector is only slightly less than the total increase in final demand. By way of contrast, industry C's final demand is assumed to change by three billion dollars, but since this industry relies only on intraindustry transactions, the total impact upon the processing sector is limited to about .9 billion dollars. The remaining additional inputs needed to satisfy the increase in final demand of three billion dollars came from outside the processing sector—in an actual underdeveloped economy a substantial fraction of these additional inputs would have to be imported. The development implications of this example are clear. If industries B, F, and D in the hypothetical economy could be expanded by stimulating the export demand for their products, industries A, E, and C would also expand as a result of the increased demand for their outputs. In planning for future development, industries similar to our hypothetical industry B would be encouraged to develop in this country. The chain reactions set off by the growth of such industries would generate expansion in other sectors of the economy. Industries like our hypothetical industry B are strongly interdependent on the input side, and such industries therefore have a high "multiplier effect" upon the rest of the economy when the demand for their products is increased.

 Since the input-output table shows only the internal structure of the underdeveloped economy, the following question might be asked: How would the leaders of the underdeveloped nation know what new industries should be added to the present economy? The answer to this question would be obtained by an examination of a similar table for an industrialized economy. Comparison of the two triangularized tables would show which new industries would draw upon the output of existing industries and sectors in the underdeveloped economy. By reference to a table for an industrialized economy policy-makers in an underdeveloped economy could estimate the total impact on their own economy of the growth of specific new industries. It would even be possible to insert in the input-output table of the underdeveloped economy rows and columns borrowed from the tables of industrialized economies. Projections would then show the total impact of the growth of the new industries represented by these rows and columns on the underdeveloped economy.39 Another type of analysis can also be made to illustrate graphically the use of input-output techniques in the study of economic development.

The "Self-Sufficiency" Chart

As indicated earlier, total self-sufficiency is not the goal of economic development, but greater self-sufficiency is. The effects of increasing self-sufficiency can be illustrated by reference to what has been called a self-sufficiency or "skyline" chart. Two hypothetical skyline or development charts are illustrated by Chart 5-1. It should be noted that these hypothetical charts are not at all realistic. They have been deliberately kept simple to facilitate description. Examples of actual skyline charts for the United States, Israel, Egypt, and Peru are given by Leontief in his September 1963 Scientific American article.40 The actual charts, which are quite detailed, show the striking differences in structure between a highly industrialized and three underdeveloped economies.

The principles involved in constructing such charts are fairly simple, although a tremendous amount of detailed analysis lies behind an actual skyline chart. The vertical scale of the chart shows the per cent of self-sufficiency. The horizontal axis measures output expressed in units of the country's currency. The width of each bar in the chart shows the relative importance of each of the sectors in the economy as a whole. The area of each bar up to the 100 per cent line shows the amount of production that would be required from each sector to satisfy the direct and indirect demands of the domestic economy if it were to achieve self-sufficiency. Added to the top of each bar is a crosshatched block which represents the direct and indirect requirements from that sector needed to produce its exports entirely from domestic resources. A crosshatched block (with lines running in the opposite direction) is then subtracted from this total. The latter shows the amount of production that would be required from this sector, directly and indirectly, to produce goods that are now imported. The heavy black line, which suggests the appearance of a skyline on the horizon, represents the actual total output of each of the sectors on the chart.

In the hypothetical charts (Chart 5-1) the industrialized economy has a surplus of "direct and indirect exports" while the underdeveloped economy has an export deficiency which is more than offset by the "direct and indirect imports." Although these charts are not based on actual data they accurately represent the situation as it exists today between an industrialized nation such as the United States and any one of the world's underdeveloped economies. Part of the surplus exports from the industrialized economy go to the underdeveloped economy, and many of these exports are designed to increase the "self-sufficiency" of the latter. As capital goods are imported by the underdeveloped economy and its internal structural interdependence is increased, its skyline chart should tend to look more like that of the industrialized economy. Indeed, one useful application of this technique is to project the changes in structure which will occur as families of structurally related industries grow in a country.41

It has been noted that the input-output technique can be used for regional as well as national development purposes. A recent study by the Mississippi Industrial and Technological Research Commission illustrates the regional application of the techniques discussed above in connection with underdeveloped nations.42 Carden and Whittington have constructed input-output tables for the state of Mississippi, using 1961 data, and from these they have derived a skyline or self-sufficiency chart. On the basis of their analysis they have identified a number of structurally related industries which should be encouraged to grow in the state if it is to optimize the use of its resources and significantly increase its per capita income. As in the case of national development programs, these authors do not suggest that Mississippi should become completely self-sufficient. They define self-sufficiency as "the amount of exports which each industry is able to generate to pay for the imports of goods of that industry which the economy does not produce."43 They recognize fully the benefits of specialization and exchange. There is a great deal of specialization within "industries," however, so that "self-sufficiency" as they have defined it would mean not less but more interregional trade. As in our earlier hypothetical analysis, their study indicates the benefits that would accrue to the state if industries which would derive many of their inputs from other Mississippi sectors could be encouraged to expand in the state.


The input-output method is now being used as a basic analytical tool by government agencies in a large number of countries. While Leontief's original model was applied to a private-enterprise economy — in which changes in final demand are autonomous it has been used extensively in recent years by countries with centrally planned economic systems. This demonstrates the value-free nature of input-output analysis. It is an analytical tool which is not tied to any particular system of economic decision-making.

The advantage of input-output analysis in the study of economic development is that it shows in detail how changes in one or more sectors of the economy will affect the total economy. No one has claimed that all that is needed is an input-output table (or any other analytical tool) to achieve economic growth. Leontief has put the case well: "The mere existence of an elaborate projection will not, of course, bring about economic growth. Much political acumen and drive, much sweat and tears goes into the actual realization even of the best-conceived developmental plan. Progress, however, will be faster along a road well mapped in advance and the cost of progress in terms of labor, capital and human sacrifice considerably less."44

Chart 5-1 (click here)


BARNA, TIBOR (ed.), Structural Interdependence and Economic Development, Proceedings of an International Conference on Input-Output Techniques, Geneva, September 1961 (New York: St. Martin's Press, 1963).

_____, The Structural Interdependence of the Economy, Proceedings of an International Conference on Input-Output Analysis, Varenna, June 27-July 10, 1954 (New York and Milan: John Wiley & Sons and A. Giuffre, 1956).

Board of Trade and Central Statistical Office, Input-Output Tables for the United Kingdom, 1954 (London: Her Majesty's Stationery Office, 1961).

CARDEN, JOHN G. D. and F. B. WHITTINGTON, JR., Studies in the Economic Structure of the State of Mississippi (Jackson, Miss.: Mississippi Industrial and Technological Research Commission, 1964).

European Committee for Economic and Social Progress (CEPES), French and Other National Economic Plans for Growth (New York: Committee for Economic Development, June 1963).

KANTOROVICH, L. C., "Mathematical Methods of Organizing and Planning Production," Management Science, VI (July 1960), 366-422.

LANGE, OSKAR, Introduction to Econometrics (2d ed.; New York: The Macmillan Company, Pergamon Press, 1963).

LEONTIEF, WASS1LY, "The Decline and Rise of Soviet Economic Science," Foreign Affairs 38 (January 1960), 261-72.

_____, "The Structure of Development," Scientific American, CCIX (September 1963), 148-66.

LEVINE, HERBERT S., "Input-Output Analysis and Soviet Planning," American Economic Review, LII (May 1962), 127-37. Comments by James Blackman, pp. 159-62.

TREML, VLADIMIR G., "Economic Interrelations in the Soviet Union," Annual Economic Indicators for the U.S.S.R., Joint Economic Committee (February1964) (Washington, D.C.: U. S. Government Printing Office, 1964), pp. 185-213.



1 For an excellent discussion of the 50-sector table, see W. Duane Evans and Marvin Hoffenberg, "The Interindustry Relations Study for 1947," The Review of Economics and Statistics, XXXIV (May 1952), 97-142. See also General Explanations of the 200 Sector Tables: The 1947 Interindustry Relations Study (United States Department of Labor, BLS Report No. 33, June 1953).

2 Jerome Cornfield, W. Duane Evans, and Marvin Hoffenberg, Full Employment Patterns, 1950 (U.S. Department of Labor, Bureau of Labor Statistics, Serial No. R. 1868, 1947), reprinted from the February and March issues of the Monthly Labor Review.

3Business Week (August 29, 1953), 26.

4 Ibid.

 5 The Netherlands Economic Institute, Input-Output Relations, Proceedings of the Conference on Interindustrial Relations held at Driebergen, Holland (Leiden 1953).

 6 Tibor Barna (ed.), The Structural Interdependence of the Economy, Proceedings of an International Conference on Input-Output Analysis (New York and Milan: John Wiley & Sons, Inc., and A. Giuffre, 1956).

 7 Tibor Barna (ed.), Structural Interdependence and Economic Development, Proceedings of an International Conference on Input-Output Techniques, Geneva, September 1961 (New York: St. Martin's Press, 1963).

 8 From the preface by Wassily Leontief to Structural Interdependence and Economic Development, p. V.

 9 Felix de Clinchamps, "The Role of Private Enterprise in the Preparation of the Plan," French and Other National Economic Plans for Growth, European Committee for Economic and Social Progress (CEPES) (New York: Committee for Economic Development, 1963), p. 62.

 10 See Hollis B. Chenery and Paul G. Clark, Interindustry Economics (New York: John Wiley & Sons, Inc., 1959), pp. 251-67, and Hollis B. Chenery, Paul G. Clark, and V. Cao-Pinna, The Structure and Growth of the Italian Economy (Rome: U.S. Mutual Security Agency, 1953). See also "The ABC of Input-Output," reprinted from the (London) Economist (September 19 and 26, 1953) by St. Clements Press, Ltd., and Input-Output Tables for the United Kingdom (London: H.M.S.O., 1961).

 11 Wassily Leontief, "The Decline and Rise of Soviet Economic Science," Foreign Affairs, XXXV III (January 1960), 262.

 12 Ibid., 263.

 13 See his Introduction to Econometrics, 2nd rev. ed. (New York: The Macmillan Company, 1963), pp. 9-23. More than half of this book is devoted to input-output analysis and the related technique of linear programming.

 14 The paper was first published in Germany, but was later translated for publication in Russia. See "The Decline and Rise of Soviet Economic Science," op. cit., pp. 269. See also "Soviet Planners Bootleg Western-Style Economics," Business Week (June 13, 1959), 92-96.

 15 L. V. Kantorovich, "Mathematical Methods of Organizing and Planning. Production," Management Science, VI (July1960), 366-422 (translated by Robert W. Campbell and H. W. Marlow).

 16 Joint Economic Committee, Congress of the United States, Annual Economic Indicators for the U.S.S.R. (Washington: U. S. Government Printing Office, 1964), pp. 185-218. This is a summary of a study "Soviet 1959 Interindustry Model: Reconstruction and Analysis," prepared for the Research Analysis Corporation, McLean, Virginia, by Dr. Vladimir G. Treml of Franklin and Marshall College.

 17 Herbert S. Levine, "Input-Output Analysis and Soviet Planning," American Economic Review, LII (May 1962), 127-37.

 18 Joint Economic Committee, op. cit.

 19 For a more complete discussion see the lucid presentation by Levine, op. cit., pp. 128-31.

 20 Ibid., p. 130.

 21 Ibid., p. 132.

 22 Ibid., p. 133.

 23 Treml, op. cit., p. 192.

 24 Ibid., p. 185.

 25 Ibid., p. 188.

 26 Ibid.

 27 Cf. Lange, op. cit., pp. 214-24, and Treml, op. cit., p. 189.

 28 Tibor Barna (ed.), Structural Interdependence and Economic Development (New York: St. Martin's Press, 1963), p. 2

 29 Barna, op. cit., p. 6. For a different point of view on this position see the comments by Walter Isard in Input-Output Analysis: An Appraisal (Princeton: Princeton University Press, 1955), pp. 366-67.

 30 See for example W. W. Rostow (ed.), The Economics of Take-Off into Sustained Growth, Proceedings of a Conference held by the International Economic Association (New York: St. Martin's Press, 1963).

 31 See Allen G. B. Fisher, The Clash of Progress and Security (London: Macmillan and Co., Ltd., 1935), and Colin Clark, The Conditions of Economic Progress (London: Macmillan and Co., Ltd., 1940).

 32 This is true of the analysis of "underdeveloped" regions as well as of industrialized nations.

 33 Wassily Leontief, "The Structure of Development," Scientific American, CCI X, No. 3 (September 1963), 148-66.

 34 See Structural Interdependence and Economic Development, especially parts I—III. For a related approach which, however, differs in a number of respects from the basic input-output model, see Leif Johansen, A Multi-Sectoral Study of Economic Growth (Amsterdam: North-Holland Publishing Company, 1960).

 35 Leontief, op. cit., p. 159.

 36 For an example of the triangularized input-output table for an actual economy (that of Israel) see Leontief, op. cit., pp. 152-53: a graphical comparison of triangularized input-output tables for the United States and the OEEC nations of Western Europe is given on pp. 150-51.

 37 The details of the final demand and payment sectors in this table would be no different from those of Table 2-1, and they have been omitted here to simplify the exposition. It should be remembered, however, that among the purchases not shown in Tables 5-1 and 5-2 are those from other countries.

 38 In reading a triangularized input-output table it is useful to recall that the industries below any given row (say row D in the example used here) are that industry's suppliers while the industries above that row are its customers. Cf. Leontief, op. cit., p. 153.

 39 This would be done by constructing a new table of technical coefficients similar to Table 5-3 using selected coefficients from an industrialized economy for the "dummy" rows and columns.

40 Pp. 162-63.

 41 For an example of a projected skyline chart superimposed on an actual chart (for Peru) see Leontief, op. cit., p. 164.

 42 John G. D. Carden and F. B. Whittington, Jr., Studies in the Economic Structure of the State of Mississippi, I (Jackson, Miss.: Mississippi Industrial and Technological Research Commission, 1964).

43 Ibid., p. 16.

 44 0p. cit., p. 166.






Regional Research Institute, West Virginia University