Contributed by: Bruno Cassiman

Incentives to Innovate

Innovation is becoming central to the strategy of firms in order to secure a sustainable competitive advantage. In this PhD level seminar we attempt to uncover the fundamental drivers of a firm’s incentive to innovate. The empirical readings present some empirical regularities on this relation while the theoretical papers present a theory (or contrast different theories) of what might be underlying this incentive. The seminar consists in understanding the potential drivers of the incentive for innovation, learning the existing theories and discussing how well these theories match the empirical findings of this multifaceted phenomenon. We start with a basic set-up which goes back to Arrow (1962), Gilbert and Newbery (1982) and Reinganum (1983). Next, we highlight different potential drivers that affect the incentives to innovate:
  • Appropriability and Spillovers
  • Complementarity of (Innovation) Activities and Assets
  • Organization and Innovation
  • Market Structure and Competition

The Basics: Incentives for R&D and Innovation

First we look at the early applications of Industrial Organization theory to R&D investments and innovation. The key starting point for this discussion is the observation that incumbents and new entrants, or, large and small firms differ in how innovative they are. Evidence of this is provided in Acs and Audretsch (1988).

We discuss the replacement effect where a monopolist has a weaker incentive to invest in R&D compared to firms in a competitive environment (Arrow, 1962) and contrast this with the business stealing effect where an incumbent has a stronger incentive to invest in R&D compared to a potential new entrant (Gilbert and Newbery, 1982). We also contrast the “economic of incentives for R&D and innovation” with alternative explanations that the management literature has provided such as inefficient investments by incumbents (Christensen, 1997) or asymmetric incentives to invest in R&D and innovation due to organizational rigidities (Clark and Henderson, 1993).

The discussion leads to the realization that incumbents and new entrants differ not only in their incentives to innovate but also in the type of innovations that they generate (more or less drastic, radical, disruptive, basic, general,…) Unfortunately, we lack a clear definition for the way these innovations “differ”. This is where careful (new) theoretical modeling with clear definitions about R&D and innovation can help.

Acs and Audretsch (1988) set up this discussion by providing evidence that small firms seem more innovative than large firms.

The reading by Gilbert (2006) gives a clear basic treatment of the different theoretical models related to the incentives to invest in R&D going from Arrow (1962) to Gilbert and Newbery (1982) and Reinganum (1983). A comparable treatment can be found in Tirole (1988).

Christensen (1997) explains the fact that incumbents have a hard time coming up with disruptive innovations because of their excessive focus on their large existing clients, suggesting that (incumbent) firms make inefficient investments in R&D and innovation. It is interesting to contrast this explanation with the previously discussed theoretical models.

Henderson and Clark (1990) argue that incumbents differ from new entrants in the constraints that their existing organization (and history) puts on the type of innovations developed. Again, this paper is interesting to contrast the different explanations for the divergence between established firms and new entrants with the developed theoretical models. This paper provides an explanation for the asymmetric incentive to invest in certain technologies.

In the following discussion we go on to look at other elements of the firm environment that can affect the incentive to invest in R&D and innovation.

Appropriability and Spillovers

Arrow (1962) early on pointed out the fact that unlike regular investment incentives the incentives to invest in R&D and innovation are affected by the fact that not all returns to the investment can be appropriated, the so-called spillovers. Information leaking out to other market participants reduces the firm’s incentive to invest in R&D and innovation. Jaffe (1986) provides evidence about the importance of these spillovers at the firm level affecting innovation, profits and market value. D’Aspremont and Jacquemin (1988) captures the essence of this spillover effect in a very simple and stylized model which allows us to be precise about the effects of spillovers and possible remedies such as the organization of research joint ventures.

Complementarity of (Innovation) Activities and Assets

So how can firms now profit from innovation? Teece (1986) in his seminal paper argues that firms can profit from innovation by either selling their innovation on the technology market or incorporating their innovation in a product or process and commercializing the idea. The decision depends on the complementary downstream assets that the firm controls. This discussion leads us to realize that the marginal returns to innovation depend on other related activities of the firm.

Milgrom and Roberts (1995) lays out the theory of complementarity between different activities of the firm – i.e. the fact that the marginal return to one activity depends on the level of another activity. They apply this to the choice between flexible versus standardized manufacturing technology and demonstrate that the choice of production technology has important consequences for many other – complementary – decisions of the firm.

Cassiman and Veugelers (2006) apply this to the innovation environment where they show that own R&D investments and external knowledge acquisition are complementary innovation activities affecting the incentive to invest in innovation.

Organization and Innovation

How companies organize internally can seriously affect incentives to invest in R&D and innovation. Saxenian (1994) compares how HP and Digital Equipment Company were organized in the early 90s. The companies were very similar in size and active in the same industries, but their internal organization differed radically, which lead to important differences in their innovation outcomes. Aghion and Tirole (1994) provide a simple theoretical model that seems to capture these important effects of organization on R&D incentives by relating ownership of the innovation to the effort that researchers or other important agents expend in order to commercialize the innovation and appropriate returns from this innovation. Lerner and Mergers (1997) attempts to test the Aghion and Tirole (1994) model and discuss the relative importance of different elements in biotech alliances on innovation outcomes.

Market Structure and Competition

Finally, the relation between market structure and the incentives to invest in R&D and innovation is probably the most studied and debated in Economics. Cohen and Levin (1989) lay out the key empirical findings on this relationship. Vives (2008) is an attempt to integrate the disparate theoretical models that have been developed in the literature. Nevertheless, there is still little consensus on the effects of competition on the incentives to invest in R&D and innovation. As Vives (2008) shows, it depends on how competition is actually defined and on the type of innovation considered.

From this discussion on the drivers of the incentive to invest in R&D and innovation it becomes very clear that we still do not have theoretical models that predict sufficiently well the observed phenomenon. Many interesting questions (fortunately for researchers!) remain unresolved.