The Causal Effects of Competition on Innovation: Experimental Evidence
GAICIL Insights
Economists have argued for decades over whether competition spurs or smothers innovation.1 This paper settles part of the fight in a laboratory. The Causal Effects of Competition on Innovation: Experimental Evidence, by Philippe Aghion, Stefan Bechtold, Lea Cassar and Holger Herz, sidesteps the endogeneity that plagues field data, where competition and innovation influence each other. Two experiments analyze the causal effects of competition on step-by-step innovation, where costly R&D moves technology up one step.
Competition is measured inversely by the ex post rents of firms operating at the same technological level, so-called neck-and-neck firms. The results are clean. Increased competition significantly raises R&D investment by neck-and-neck firms, strong evidence of an escape-competition effect. It decreases R&D by firms lagging behind, particularly when the time horizon is short. And it changes industry composition, reducing the fraction of sectors where firms are neck and neck while the average technology level of leading firms rises. Rivals running side by side sprint; those far behind slow down.
All results match step-by-step innovation models. The lab offers distinct advantages. It addresses endogeneity directly, so results capture causal effects. It disentangles competition's effects in leveled and unleveled sectors. It varies the time horizon, showing that competition's discouraging effect on laggards is stronger in the short horizon, which suggests an anticipated escape-competition effect operates in the long horizon. Give laggards a longer future and hope returns.
The method has wide reach. Experiments could study patent protection, R&D subsidies, intellectual property legislation, antitrust policies and contractual or institutional arrangements. Industrial organization and law and economics often identify counteracting effects without saying when each dominates, and experiments can sort out such debates.
The parallel with today's technology is close. Artificial intelligence is a live demonstration. AI markets feature neck-and-neck competition among leading developers and laggards trying to catch up. Intense competition among firms at similar technological levels spurs innovation as each seeks to pull ahead. That dynamic helps explain the rapid pace of progress when several AI developers compete closely at the frontier. A frontier with several close rivals is a frontier in a hurry.
Laggards are the qualification. Competition can discourage firms far behind, especially with short horizons. AI policy should attend to conditions under which laggards can catch up. Longer horizons, letting firms anticipate the benefits of eventually reaching the frontier, may sustain their incentives. Policies supporting knowledge diffusion and lowering catch-up costs keep more firms in the race. Open research and accessible tools are ways of lengthening a laggard's horizon.
Industry composition shifts too. As competition rises, sectors become less likely to be neck and neck and leaders' technology rises. Competition drives frontier progress while shaping market structure, and policymakers should recognize that competitive pressure can produce leadership gaps as a byproduct of innovation. A widening gap can be a symptom of healthy rivalry, so its cause needs checking before it is treated as abuse.
The method helps AI governance directly. Many AI policy questions involve counteracting effects, and laboratory experiments can isolate them, complementing field evidence.
For competition policy, the results provide causal support for the view that competition among close rivals spurs innovation. Preserving neck-and-neck competition in AI markets, where several firms vie for leadership, may be particularly important. Policies preventing any single firm from escaping competition permanently through anticompetitive conduct help maintain the dynamic.
Read as a whole, competition's effect on innovation depends on firms' technological positions. Neck-and-neck firms innovate more under competition; laggards may innovate less, especially with short horizons. As AI markets evolve, these dynamics can guide competition policies that sustain frontier innovation while keeping the path open for challengers.
A hard question for AI antitrust: the market structure to defend is a close race at the top, with fragmentation a separate question. Keep several leaders neck and neck, and give the rest a long enough horizon to believe they can join them.
Introduction
We design two laboratory experiments to analyze the causal effects of competition on step-by-step innovation. Innovations result from costly R&D investments and move technology up one step. Competition is inversely measured by the ex post rents for firms that operate at the same technological level, that is, for neck-and-neck firms. First, we find that increased competition leads to a significant increase in R&D investments by neck-and-neck firms. Second, increased competition decreases R&D investments by firms that are lagging behind, in particular if the time horizon is short. Third, we find that increased competition affects industry composition by reducing the fraction of sectors where firms are neck-and-neck. All these results are consistent with the predictions of step-by-step innovation models.
Takeaway
In this paper, we provided a first attempt at analyzing the effect of competition on step-by-step innovation in the laboratory. Using the lab instead of field data has several advantages. First, it addresses the endogeneity issue head on: our results do capture causal effects of competition on innovation. Second, the lab experiment allows us to disentangle the effects of competition on innovation in leveled and unleveled sectors. In particular, we find strong evidence of an escape-competition effect in neck-and-neck sectors. Third, our design allows us to study how the effect of competition on innovation varies with the time horizon. Consistent with theory, we show that the Schumpeterian effect is stronger in the short horizon treatment than in long horizon treatment, suggesting that in the latter case, an anticipated escapecompetition effect is also at work. Fourth, we are able to identify the causal effect of competition on industry composition and on aggregate outcomes. We find that, as competition increases, sectors become less likely to be neck and neck, and the average technology level of the leading firm increases. The methodology used in the paper can be used to sort out other open debates in industrial organization and law & economics. For example, one could use experiments to study the effects of patent protection or R&D subsidies, or the relative performance of various intellectual property legislations or the impact of various antitrust policies, or the effects of various contractual or institutional arrangements, on innovation and entry.1
The industrial organization and law & economics literatures often point to counteracting effects without always spelling out the circumstances under which one particular effect should be expected to dominate. We believe that lab experiments can fill this gap by providing more precise predictions as to when such or such effect should indeed dominate. This and other extensions of our analysis in this paper are left to future research.
In Acemoglu and Akcigit (2012), this corresponds to our parameter h.2 Experimental testing may be particularly useful given the homogeneity of intellectual property regimes across countries resulting from international harmonization efforts.
Notes
- Philippe Aghion et al., The Causal Effects of Competition on Innovation: Experimental Evidence, 34 J.L. Econ. & Org. 162 (2018), https://doi.org/10.1093/jleo/ewy004. Takeaway source: Conclusion, reproduced from a free author or repository copy (working-paper version). ↩
- Acemoglu, D. and U. Akcigit (2012). Intellectual Property Rights: Policy, Competition and Innovation. Journal of the European Economic Association 10, 1–42.. ↩
Cite as: Philippe Aghion et al., The Causal Effects of Competition on Innovation: Experimental Evidence, 34 J.L. Econ. & Org. 162 (2018), https://doi.org/10.1093/jleo/ewy004
Header image: Data center backup generators.jpg by Rsparks3, CC0, via Wikimedia Commons.
