In this chapter, I use a theory of research and development (R&D) rivalry to explain why an industrial firm would diversify its R&D efforts. The theory is then used throughout Parts III and IV to analyze business strategy and public policy.

Introduction

This chapter's explanation for diversified R&D provides the theoretical underpinning of the observations, in Chapters 9 and 10, about purposive diversification. Further, the formal model of homogeneous R&D rivalry describes the intense form of R&D competition which Chapter 11 suggests may be so unattractive from a private standpoint that a firm, when faced with numerous competitors, would deliberately face the risks of thoroughly new and untried approaches to the R&D problem. Additionally, I shall use the model to explain the social economic welfare consequences for R&D investment in the alternative market structures of monopoly or completely cooperative R&D ventures, Nash noncooperative R&D equilibria, and the free-entry Nash noncooperative equilibrium. Understanding the social economic welfare implications of these different market structures will provide the theoretical backdrop for our discussion of industrial policy in Part IV.

The model that I shall use in this chapter (and throughout the discussions of R&D in the remaining chapters) is essentially Lee and Wilde's (1980) reformulation of Loury's (1979) model, which in turn is an extension of the ideas of Scherer (1967b), Barzel (1968), and Kamien and Schwartz (1982). The model has been a popular and well-accepted one; Hartwick (1991) provides an example of its use to study problems in industrial organization.

Numerous government policies attempt to spur spending for research and development (R&D) in order to achieve socially desirable innovations and productivity growth. Two government policies for promoting industrial R&D that have received a considerable amount of attention recently are the granting of tax reductions in return for increased R&D spending and the lessening of potential antitrust liabilities faced by cooperative R&D ventures. The effectiveness of both policies has been questioned. In this chapter, I use the theory developed in Chapter 8 to show that in circumstances for which the policies have been expected to work they are likely to fail. I then introduce two novel tax policies, either of which can in theory induce socially optimal R&D investment when the conventional policies would fail. The policies are designed to mimic socially desirable competitive pressure in circumstances under which it would otherwise be absent. A policy that would simulate competitive pressures could make an especially important contribution now because of the growing trend toward cooperation among previously rivalrous firms. Thus, I shall consider the practical possibilities for the theoretical taxes proposed.

Introduction

Chapter 12 described circumstances in which cooperation among an industry's firms are conventionally considered socially desirable and for which free-entry noncooperative rivalry is often presumed undesirable. These same circumstances have often been presumed to justify tax breaks in order to stimulate R&D.

In the autumn of 1980, having been granted a year-long leave of absence from Dartmouth College, I arrived in Washington, D.C., at the offices of the United States Federal Trade Commission Line of Business (FTC LB) Program. As a visiting research economist, I was granted access to the program's confidential data describing in unprecedented detail the diversification of the United States' largest manufacturing firms. After roughly a decade of working, first as an in-house economist and then as an outside consultant, with the FTC's remarkable data. I decided to bring together my observations about the FTC LB reporters' diversified activities. This book is the result. In the book, I introduce and apply methodology that discerns groups of manufacturing industries that are related because of complementarities in production, marketing, distribution, and research and development (R&D) activities. Manufacturing firms purposively diversify to exploit such complementarities, and I explore hypotheses about that behavior – i.e., purposive diversification – and ensuing economic performance. The book studies product diversification's effects on both static allocative efficiency and the optimality of R&D investment.

The study of those hypotheses about purposive diversification yields new perspectives on the policy debate about cooperation versus competition among firms. The debate is an old one that has flared anew. Chernow (1990, p. III) gives an interesting perspective on the debate at the beginning of the twentieth century when the firms involved were the railroads, the steelmakers, and the oil refiners.

In this chapter, I shall use the ideas and findings from the preceding chapters to compare the diversification of R&D in the mid-1970s with the mid-1980s cooperative R&D that is protected by the National Cooperative Research Act of 1984 (NCRA, U.S. 98th Congress, October 1984). The evidence suggests that the effect of the law may well be to reduce innovative investment and cause investment to be further away from the socially optimal amount. Yet the NCRA model has been touted by business executives, economists, and policymakers as a way to improve U.S. industrial performance, and has become the basis for new law encouraging production joint ventures, as we shall see in Chapter 13.

I shall ask whether the initial response to the NCRA suggests that the act stimulated the types of activity envisioned by those supporting the new law. With the exception of a few examples, I use the first year and a half of data about the coooperative ventures registering under the new law because that seemed a large enough amount of information about initial reaction to gain some insight into whether the behavior reflected the concerns of the law's architects. I explore the issue in its historical context. For example, did the NCRA stimulate R&D in industries where the productivity slump hit hard, as the arguments made in support of the legislation suggested that it would? Of course, the desire to stimulate R&D in industries subject to a productivity slump may have been misguided.

In Chapter 6 we saw that traditional views of the link from structure to performance explain only a very small portion of the systematic variance (the variance explained by firm and industry effects) in profits across firms and industries. In other words, we saw that firms differ significantly and that industries differ significantly, yet our traditional models do not explain very much of those significant differences across firms and industries. That finding, as presented in Scott and Pascoe (1986), has attracted some attention in the literature and has been replicated with very different data (Amato and Wilder, 1990). Here in Chapter 7 we shall document the importance of firm and industry effects in R&D intensity and see that the often postulated link from seller concentration to research and development (R&D) intensity is present in the Federal Trade Commission Line of Business (FTC LB) data, but that it disappears once firm and industry effects are controlled. Further, the effect of seller concentration is but a small part of the systematic variance in R&D intensity across business units.

Scherer (1965, 1967a) pioneered the result that technological opportunity is far more important for understanding variance in innovative activity than are measures of rivalry. The result was documented in the FTC LB data by Scott (1984) and confirmed by Levin et al. (1985) using a different procedure with those data.

Large manufacturing firms typically operate in many markets. As a result, when we analyze a market's performance as a function of its structure, we need to consider the diversification of the market's sellers and their multimarket contact. The operations of the typical firm among the largest 1000 U.S. manufacturers span several lines of business. Table 1.1 provides the frequency distribution describing the number of manufacturing lines of business (LBs – an LB is a company's operations in a particular industry) for the 437 companies in the sample used in this chapter. The companies are all among the largest 1000 U.S. manufacturers and comprise the Federal Trade Commission Line of Business (FTC LB) Program sample for 1974. These large companies average a bit under eight manufacturing LBs per company.

Of course, the motives for such diversification are numerous. For just a sampling of the variety of commentary through the years, consider Penrose (1959), Gort (1962), Rumelt (1974), Berry (1975), Mueller (1987), and Montgomery and Werner felt (1988). Although random discrepancies in the valuation of assets (Gort, 1969) and the risk aversion of managers (Amihud and Lev, 1981) can motivate pure conglomerate diversification, the purposive pursuit of (private) efficiencies because of complementarities across industry categories motivates “related” diversification. I shall focus on the trade-off that such diversification creates between the possibilities for technical efficiencies and for market power and then investigate the ensuing overall performance effects on static and dynamic efficiency.

I shall explore the causes and effects of diversification.

In Chapter 1, we observed that diversification can be a way to achieve economies of scope, and the pursuit of such economies could explain why, for example, diversified firms combine lines of business (LBs) that share distribution channels. However, we also noted that there are hypotheses outstanding, suggesting that diversification can create private but not necessarily social gains if it is also a means to achieve market power. In this chapter, we shall begin our exploration of that possibility by developing the hypotheses; in subsequent chapters we shall examine evidence about the hypotheses.

Introduction

The Celler-Kef auver Act of 1950 amended Section 7 of the Clayton Act and redirected mergers toward the conglomerate variety (Scherer, 1980, pp. 123–124). Do the ensuing conglomerate mergers reflect a redirecting of mergers for market power? Since firms' limited organizational capacity for assimilating acquisitions could no longer be focused on horizontal mergers, perhaps conglomerate mergers afforded an alternative source of market power. Yet, as Section 2.2 explains, even the activist antitrust enforcement policies of the 1960s did not address the issue fully, and recent merger enforcement has certainly ignored the possibility.

After Section 2.2 discusses the theories about market power that have been used in policy toward conglomerate mergers, Section 2.3 turns to a theory not used in such public policy – namely the theory that multimarket contact can increase market power. Section 2.4 concludes by emphasizing the complementarities among the older and newer theories about multimarket contact.

This chapter explores an important possibility suggested by the findings of Chapter 7 and Chapter 9 taken together. Chapter 7 showed that the correlation between seller concentration in standard industry categories and research and development (R&D) intensity may not reflect the Schumpeterian hypotheses that market power begets R&D activity as commonly supposed. Chapter 9 showed that purposive diversification of R&D occurs and shapes R&D behavior and productivity performance. Might it not be, then, that within the universe of the very largest firms the state of competition should be evaluated not for standard industry categories but instead for the multimarket groups of related categories and the firms within them whose operations span the set of categories in the group?

Introduction

Many economists have been intrigued by Schumpeter's (1942) vision of technological progress driven by firms fighting to survive competition from new products, new processes, or new organizational forms. But his proposition that the firms at the heart of that potentially socially beneficial struggle are necessarily large and monopolistic – that firm size and monopolistic power promote technological advance – has not found strong empirical support. Focusing on the big facts about how technological change has benefited society, Schumpeter argued that any static allocative and technical inefficiencies caused by monopolies would be overwhelmed by the good effects of the desirable technological change that the monopolies induce. Schumpeter's hypothesis is of course distinct from the empirical tests used to explore it.