Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Macroeconomics and Microeconom~cs of Innovation: The Role of the Technological Environment . .. . . ANN F. FRIEDLAENDER If America is to continue to grow and thrive, the answer to the question of how productivity growth and technical change take place is crucial. The answer can perhaps only be obtained by using analytic frameworks that confine and synthesize the insights of technologists, . , . engineers, ana economists. As an economist who has worked with engineers and technologists for many years, I am struck by the fundamental lack of communication that exists between them in discussing the sources and nature of innovation and productivity growth. This anses, I believe, because of basic differences in the way each group visualizes the production process and the nature of innovation. Since each group brings a very different perspective to the prob- lem, increased communication should lead to a richer analytic framework and a deeper understanding of the sources and nature of technological change. OVERVIEW A brief description of the approach of each group will indicate some of the ways in which a greater synthesis of the two approaches could increase our understanding of technical change and productivity growth. While something of a cancature, it is probably fair to state that economists view the production function as an abstraction, a "black box" if you will, that permits the transformation of certain inputs (labor, capital, materials, and so forth) into different outputs; Hey have little, if any, interest in the inner workings of that black box. Thus, to economists, the production func- 327
328 ANN F. FRIEDLAENDER tion is merely that a general functional expression relating outputs to inputs; they have little regard for the specific functional fonn or how the specific technology employed may affect either the functional form or the parameters of the production function. Indeed, it is only within the past decade that economists have generally realized that the specific functional form used to describe technology may impose severe restrictions on its assumed nature (e.g., the unitary elasticity of substitution associated with the Cobb-Douglas production function) and have adopted so-called flexible functional foes in analyzing production relationships and productivity growth. In contrast, engineers and technologists have little interest in the abstract "production function" as such; instead, they are interested in describing the actual technical relationships that permit the translation of specific amounts of labor, machines, and materials into widgets or other commodities. Thus, engineering production functions tend to be very detailed and to focus on He technical environment in which production takes place. Consequently, although relative input prices drive the allocation of inputs within the context of the economists' production or cost functions, Hey tend to be ignored in engineering production functions. Conversely, while He technological en- vironment tends to drive engineering production functions, it is often ignored in economic production functions. Thus, by focusing on He technological environment, engineers and technologists can provide economists with im- portant insights Hat could enrich the economists' framework and analytic models. IMPORTANCE OF THE TECHNOLOGICAL ENVIRONMENT A few examples will illustrate the kinds of things that I am refemng to. Consider the question of productivity grown in the automobile industry. In recent years, there has been considerable concern about the level of pro- duciivity in the American automobile industry relative to that in Japan. In particular, the concern has focused on two organizational differences between He two countnes: the use of "quality-control circles" in the organization of Japanese labor and He use of "just-in-time" inventories in the Japanese production process. The use of the former presumably makes labor more productive, and the use of the latter pennits considerable savings on capital that is tied up in inventories. Suppose American producers adopted both of these innovations tomorrow and that ~ years hence an econome~ician attempted to measure productivity growth in the U.S. automobile industry. What would be He likely outcome of this econometric investigation? To see this, consider how each innovation would appear within the context of the abstract production function. Since the introduction of the quality-control circles would presumably permit more output for given amounts of labor and increase the marginal product of labor
ROLE OF TECHNOLOGIC~ E~IRONME~ IN ECONOMICS OF INNOVATION 329 relative to that of capital, it would show up as a labor-saving innovation. In contrast, since the introduction of just-in-time inventories would lead to lower capital requirements, it would show up as a capital-saving innovation. On balance, depending on the relative magnitude of these effects, one would observe increased productivity growth and technical change that could be described as labor saving, neutral, or capital saving. Clearly, however, the ex post aggregate characterization of technological change in any of these terms is not particularly enlightening, since it does not indicate how this technical change came about and what specific innovations contributed to it. What is useful is the knowledge of the specific innovations that led to the change, or, how the workings inside the "black box" were altered. While this Is something that a few economists attempt to unravel, in general too much attention is paid to the aggregate concepts rather than to the techno- logical environment that created the changes. As another examples consider industries such as transportation or com- munications, in which the configuration and utilization of the network can have a significant impact on productive relationships and technical change. To point out the implications of this, let me discuss the way in which the introduction of network variables can enhance our understanding of produc- tive relationships and technical change. First, consider a "typical" economic production function that relates output (ton-miles, passenger-miles, message units, or some disaggregated version thereof) to generalized inputs (labor, capital, materials, and so forth). It should be clear that both the configuration and utilization of the network will have a significant impact on the way in which inputs are transformed into outputs. Suppose, for example, that Grin A utilizes a network composed of a limited number of comdors, each of which is between major urban centers and thus has high density of usage. In contrast, firm B utilizes a "hub-and-spoke" network, with a central core that services a hinterland. Finally, firm C utilizes a network composed of a large number of corridors linking small urban areas without ~ hub-and-spoke network. One can readily imagine a situation in which each find utilizes the same amount of each input, but produces very different levels of output and, ~us, has very different implied levels of productivity. Although the econ- omist may simply retort Flat this is a problem of aggregation, the problem is really deeper than that, because the configuration and utilization of the network add a dimension that make the Free production processes funda- mentally different, regardless of the specification of Output. Thus again, what is going, on inside the "black box" adds fundamental insights to our under- standing of the production process. (For a full discussion of these and related points, see Chiang et al., 1984.) As a final example, let me describe a recent study I have undertaken on Me trucking industry (Friedlaender and Bruce, 1985) to try to explain its productivity growth in the period prior to deregulation (1975 to 19791. In
330 . ANN F. FENDER this particular study, costs were estimated as a function of input prices (capital, labor, fuel, and so forth), output levels (ton-miles), shipment char- actenstics (average load, average length of haul, shipment size, and so foray, and a time trend. Thus, in addition to Me usual arguments in the cost function (output, input pnces, and a time trend to capture productivity growth), the study included variables to reflect the technological environment or the way in which inputs were transformed into outputs. In this connection, it was interesting to note that virtually all of the productivity growth and technical change could be attributed to the technological factors considered rather than to He various inputs or to the scale of output. While the measures of the technological environment used in this study were admittedly crude, their statistical significance clearly indicates that what often drives productivity is the way in which inputs are utilized and the way in which output is configured. Although changes in the technological environment could have been labeled capital- or labor-saving innovations, we are surely better off knowing that what was really driving productivity growth was the way in which shipments were distnbllted over the network. Ideally, both for purposes of policy and for purposes of understanding technical change, we would like to know the specific organizational or logistical changes that were made and how they could be expected to affect productivity so that all firms could be encouraged to undermlce them. ~ To summarize my basic point, I believe that a much richer analytic frame- work is possible if economists would attempt to incorporate the fundamental insight of engineers and technologists- that the organizational and techno- logical environment can have a major impact on productivity and technical change. While abstractions relating to the elasticities of substitution, returns to scale, and neutral or nonneutral technical change are useful, they are that much more meaningful if the interactions between input utilization and the technological environment are recognized and taken into account. COMMENTS ON CHAFERS BY JORGENSON AND MANSFIELD I am disappointed to note that in their discussions in this volume, which deal with microeconomic issues related to technical change and productivity grown, Dale W. Jorgenson and Edwin Mansfield largely ignore the potential insights of technologists and engineers with respect to innovation and pro- ductivity growth. Let me begin with Jorgenson's chapter, which analyzes the sources of productivity grown in the American economy in the postwar period. As is true in most analyses of this type, Jorgenson employs a neoclassical frame- work, which assumes that firms operate in perfectly competitive input and output markets and that they utilize an aggregate production function that transforms aggregate inputs (capital, labor, electrical energy, and nonelec-
ROLE OF TECHNOLOGICAL ENVIRONMENT IN ECONOMICS OF INNOVATION 331 trical energy) into an aggregate output. Recognizing the importance of sec- toral differences, Jorgenson utilizes a disaggregated sectoral analysis and estimates a price frontier, adjusted for productivity growth, for 35 industnes. Not surprisingly, he finds that patterns of productivity growth are not uniform, but vain widely among industries; no more than ~ of the industries exhibited common patterns of productivity growth. It is admittedly useful to know that 8 of the 35 industries exhibited a pattern of productivity growth in which increases in the prices of capital, labor, electrical energy, and nonelectrical energy retard productivity growth and in which a rise in the price of materials stimulates productivity growth. More- over, it is also useful to know that 20 of the 35 industries exhibited a pattern of capital-using productivity growth. Nevertheless, it would be even more useful to know why those industries exhibited different patterns of growth and what determined the magnitude of the stimulating or inhibiting effects of input price changes on productivity growth. Without such knowledge, neither businessmen nor policymakers have any real guides concerning ap- propriate actions they could take to increase productivity growth in particular industries and in the nation as a whole; nor do they know whether past relationships between input prices and productivity growth can be expected to continue in the future. To continue with this theme, the following ~ industries exhibited a common pattern in which price increases in all inputs (with the exception of materials) inhibited productivity growth: tobacco, textiles, apparel, lumber and wood, printing and publishing, fabricated metals, motor vehicles, and transporta- tion In contrast, the following S industries exhibited the same pattern of productivity growth, with the exception that increases in the cost of capital stimulated rather than retarded productivity growth: rubber; leather; instru- ments; gas utilities; and finance, insurance, and real estate. These lists of industries immediately bring the following questions to mind: (1) What was there in the technological environment in the 8 industries that led them to exhibit a common pattern of productivity growth? (2) What was there in the technological environment of the 5 industries that led Rem to exhibit a common pattern of productivity grown? (3) What was there in the techno- logical environment of the 13 industries that led them to exhibit common patterns of productivity growth, with the exception of capital? (4) And fi- nally, to what extent were these similarities and differences due to the nature of the output, the nature of the inputs, or the technological environment in which production took place? Thus, Jorgenson seems to skirt the basic ques- tion of why productivity growth is stimulated in some industries by rises in certain input prices and why it is inhibited by similar price increases in other industries. Whether the answers to these questions can be found in the analy- ses of engineers and technologists is unclear, but it would be interesting to get their insights into this problem.
332 ANN F. FRIEDLAENDEl? My comments on Manshleld's discussion echo this theme, although perhaps in a somewhat gentler voice, since Mansfield does attempt to focus on the following specific factors that affect innovation and productivity growth: (1) the level and nature of expenditures on R&D and basic research; (2) the role of imitation costs, imitation times, patents, bamers to entry, and in- dustnal concentration; and (3) the diffusion of innovation and technology transfer. Thus, instead of focusing on the aggregate production function, Mansfield focuses on issues that relate to the productive and technological environment. Nevertheless, one wishes that Mansfield and his fellow economists had gone further. If, for example, labor and capital productivity are positively affected by the level of R&D expenditures, one would like to know how this process takes place. Is it the existence of R&D expenditures, per se, that stimulates labor arid capital productivity, or are R&D expenditures trans- fonned into specific innovations that make labor or capital more productive? Similarly, why does basic research stimulate total factor productivity inde- pendent of the level of R&D expenditures? As Mansfield points out, the question is whether this reflects a tendency for basic research findings to be exploited more fully by finns and industries that were responsible for it, whether it reflects a tendency for applied R&D expenditures to be more effective when earned out in conjunction with some basic research; or whether basic research is merely serving as a proxy for long-tenn R&D. Again, we are left with many tantalizing questions, but little analysis to answer them. Thus, I would like to close by encouraging economists to think more seriously about and confront the issues posed by the disassembly of the black-box production function. While such a disassembly would certainly reduce He elegance of economic analysis and limit the generality of the models employed, it should also provide greater insights into the question of how productivity growth and technical change take place. If America is to continue to grow and thrive, Me answer to this question is crucial. In this regard, I suspect it can only be obtained by using analytic frameworks that combine and synthesize the insights of technologists, engineers, and economists. REFERENCES Chiang, Wang, Judy S. Friedlaender. and Ann F. Friedlaender. 1984. Output aggregation, network effects. and the remeasurement of ~uclcing technology. Review of Economics and Statistics 66(May):267- 277. Friedlacnder, Ann F., and Sharon Schur Bruce. 1985. Augmentation effects and technical change in the regulated trucking industry, 197~1979. In Studies in Transportation Economics. Andrew Daughety, ed. New York: Cambridge University Press (Forthcoming).
