Knowledge Spillovers, Innovation and Growth
GAICIL Insights
Free knowledge is never quite free.1 Collecting it requires having invested in the ability to understand it. Knowledge Spillovers, Innovation and Growth, by Philippe Aghion and Xavier Jaravel, revisits the notion of absorptive capacity introduced by Cohen and Levinthal in 1989, who showed that knowledge spillovers can induce complementarities in R&D efforts. Research and development does two things: it produces new innovations, and it builds the capacity to learn from others' innovations. The authors show this idea has rich implications for growth.
Absorptive capacity captures R&D's dual role. Firms invest in R&D to pursue new process and product innovations, and to develop and maintain broader capabilities to assimilate and exploit externally available information. Factors affecting the ease and character of learning therefore influence R&D incentives. The influence of technological opportunity and appropriability on inventive activity is mediated by a firm's capacity to recognize, assimilate and exploit external knowledge. A lab is also an antenna.
Three questions get the treatment. First, cross-country convergence and divergence: countries investing in absorptive capacity can benefit from knowledge developed elsewhere, while those that do not may fall further behind. Second, international coordination of climate change policies, where cross-border spillovers of clean technology knowledge shape policy incentives and effects. Third, openness in the production of ideas, where access to external knowledge and the capacity to absorb it determine openness's benefits.
The authors argue absorptive capacity sets an agenda for new empirical and theoretical work on R&D spillovers in innovation and growth. How firms and countries build the capacity to learn from others is central to how knowledge diffuses and growth spreads.
Now the AI angle. AI makes the point vivid. AI knowledge spreads rapidly through publications, open models and the movement of researchers. Benefiting from it depends on absorptive capacity. Firms and countries investing in AI research and skills can assimilate and exploit advances made elsewhere. Those that do not may struggle even when knowledge is freely available. An open model is a gift only to those able to unwrap it.
Convergence hangs in the balance. AI could accelerate convergence if developing countries build absorptive capacity, or widen gaps if they do not. Investment in education, research and technical skills is essential for countries seeking to benefit. Absorptive capacity determines whether global spillovers become local growth. The download is the easy part.
Openness and capacity are partners. Open access to AI knowledge, through published research and open models, creates potential spillovers. Realizing them requires the capacity to absorb. Combining openness with investment in absorptive capacity may maximize AI knowledge's benefits for innovation and growth. Openness without capacity helps those who are already strong.
The climate application shows spillovers shaping international policy. Clean technology knowledge crosses borders, affecting each country's incentive to invest. AI may play a similar role, with knowledge developed in one country benefiting others. International coordination on AI research and governance may need to account for these spillovers.
For firms, R&D spending has value beyond its direct outputs. Investing in AI capabilities lets firms recognize and exploit external advances. In a field moving as fast as AI, this capacity may matter as much as proprietary innovation. Cutting the research team to save money may cut the firm off from everyone else's research too.
Seen whole, the paper's message is plain: knowledge's benefits depend on the capacity to absorb it. R&D builds this capacity, creating complementarities between firms' own research and external knowledge. As AI knowledge spreads globally, absorptive capacity will determine which firms and countries benefit. A decades-old idea remains powerful for understanding convergence, climate policy and openness in the age of AI.
Companies planning to "just use" frontier AI without building in-house expertise might start here: on this framework, the firms that gain most from others' breakthroughs are the ones still doing research of their own. Absorption is an investment, and skipping it leaves the knowledge on the table.
Introduction
Cohen and Levinthal (1989) introduced the notion of absorptive capacity and demonstrated that knowledge spillovers can induce complementarities in R&D efforts. We show that this idea has rich implications when analysing important aspects of the growth process such as cross-country convergence and divergence, the international co-ordination of climate change policies, and the role of openness in the production of ideas. We also show that the notion of absorptive capacity sets an agenda for new empirical and theoretical analyses of the role of R&D spillovers in innovation and growth. Cohen, W.M. and Levinthal, D.A. (1989). ‘Innovation and learning: the two faces of R & D’, Economic Journal, vol. 99(397), pp. 569–96.
Takeaway
We have argued that firms invest in R&D not only to pursue directly new process and product innovation, but also to develop and maintain their broader capabilities to assimilate and exploit externally available information. Recognition of the dual role of R&D suggests that factors that affect the character and ease of learning will affect firms' incentives to conduct R&D. We hypothesised that such factors include the degree to which knowledge is targeted to a firm's needs, and, more generally, the character of knowledge within each of the scientific and technological fields upon which innovation depends. Our analytic model suggests these factors should both exercise a direct effect on inventive activity, and condition the influence of more conventionally considered determinants. Aside from market demand, applied economists have identified technological opportunity and appropriability as the principal industry-level determinants of the firm's inventive activity (Cohen and Levin, I989). We suggest, however, that since the effect of both of these classes of variables depend importantly upon the assimilation of external knowledge, their influence is mediated by the firm's capacity to recognise, assimilate, and exploit information. Therefore, those variables that affect the ease and character of learning should, in turn, influence how technological opportunity and appropriability conditions influence R&D spending. The empirical results generally confirm that the influence of both appropriability and technological opportunity conditions is affected by determinants of the ease of learning, particularly the targeted quality of knowledge inputs.1
This, in turn, suggests that the characteristics of knowledge that affect the ease of firm learning may represent an important class of determinants of R&D investment. The observation that R&D creates a capacity to assimilate and exploit new knowledge sheds new light on a range of questions. For example, it provides a ready explanation of why some firms may invest in basic research even when the preponderance of findings spill out into the public domain. Specifically, firms may conduct basic research less for particular results than to be able to identify and exploit potentially useful scientific and technological knowledge generated by universities or government laboratories, and thereby gain a firstmover advantage in exploiting new technologies. Likewise, basic research may permit firms to act as a rapid second mover in the face of spillovers from a competitor's innovation. This perspective implies that variables that influence the firm's incentives to learn should affect the incentives to conduct basic research. For example, as automobile manufacturing depends increasingly on fields that draw heavily on basic science, such as microelectronics and ceramics, we expect that manufacturers will expand their basic research efforts in physics and chemistry to evaluate and exploit new findings in these areas.
Notes
- Philippe Aghion & Xavier Jaravel, Knowledge Spillovers, Innovation and Growth, 125 Econ. J. 533 (2015), https://doi.org/10.1111/ecoj.12199. Takeaway source: CONCLUSION, reproduced from the openly licensed version (CC BY-NC). ↩
Cite as: Philippe Aghion & Xavier Jaravel, Knowledge Spillovers, Innovation and Growth, 125 Econ. J. 533 (2015), https://doi.org/10.1111/ecoj.12199
Header image: Scientist prepares samples in laboratory using various tools and equipment.jpg by Shixart1985, CC BY 2.0, via Wikimedia Commons.
