Competition and Innovation: An Inverted-U Relationship
GAICIL Insights
Is more competition always better for innovation?1 The answer that reshaped a debate is: up to a point. Competition and Innovation: An Inverted-U Relationship, by Philippe Aghion, Nick Bloom, Richard Blundell, Rachel Griffith and Peter Howitt, finds strong panel-data evidence of an inverted-U relationship between product market competition and innovation. Innovation rises with competition at low levels and falls at high levels.
The model has firms innovating step by step, with technological leaders and followers both doing R&D. Competition works two ways. It may raise the incremental profit from innovating for firms neck and neck with rivals, the escape-competition effect. It may cut innovation incentives for laggards, which have less to gain from catching up when competition erodes post-innovation rents. Rivalry energizes the close race and demoralizes the stragglers.
The inverted U comes from these effects interacting with equilibrium industry structure. At low initial competition, escape-competition dominates and competition encourages innovation. At higher levels, the laggard effect dominates and competition discourages innovation.
Extra predictions find support. The average technological distance between leaders and followers increases with competition, so the equilibrium degree of neck-and-neckness decreases. The inverted U is steeper when industries are more neck and neck. These confirmations strengthen confidence in the mechanism. A theory that predicts its own side effects correctly earns trust.
Here is where the paper meets the present. Artificial intelligence markets vary in competitive intensity, and the competition-innovation relationship there matters for policy. Moderate competition may maximize innovation. Too little competition may weaken incentives; extremely intense competition could discourage laggards. The useful question for AI markets is where they sit on the curve.
Escape competition describes the AI frontier vividly. When leading firms are close rivals, competition drives them to innovate to stay ahead, a dynamic visible where several firms compete closely in developing advanced models. Neck-and-neck competition can be a powerful engine of innovation. Each release from one leader is a reason for the next from another.
Laggards matter too. Firms far behind the frontier may have weaker incentives when competition is intense. In AI, where leaders may hold large advantages, policies helping laggards catch up or keeping markets contestable can sustain innovation across the industry. Without them, the field may thin to a handful of runners.
The technological-gap prediction informs market monitoring. As competition rises, gaps between leaders and followers widen. Tracking these gaps in AI markets can show whether competition supports broad-based innovation or concentrates it among a few leaders. A widening gap is a signal worth reading carefully.
Competition policy needs nuance. The relationship is not monotonic. Policymakers should assess where a market lies on the inverted U and how competition affects leaders and laggards. Increasing competition in highly concentrated AI markets may encourage innovation, while effects in already competitive markets require careful analysis. Blanket slogans in either direction fail the evidence.
The empirical approach is a strength. Testing a nonlinear relationship and several model predictions on panel data gives robust evidence on a long-debated question, evidence that has shaped thinking across economics and policy.
Step-by-step structure captures a realistic feature of innovation: firms advance incrementally, and a leader must keep innovating to hold its position. In fast-moving AI markets, this incremental race describes how leading developers release successive improvements to stay ahead of close rivals.
At bottom, competition and innovation follow an inverted U. Moderate competition, especially among neck-and-neck firms, spurs innovation through escape competition, while very intense competition may discourage laggards. As AI markets evolve, this insight can help policymakers foster the competitive conditions that maximize innovation in a technology central to future growth.
One question to put to AI antitrust debates: the sharper question is which side of the peak a market sits on. A concentrated market and a cutthroat one can both underinvest, for opposite reasons.
Introduction
This paper investigates the relationship between product market competition and innovation. We find strong evidence of an inverted-U relationship using panel data. We develop a model where competition discourages laggard firms from innovating but encourages neck-and-neck firms to innovate. Together with the effect of competition on the equilibrium industry structure, these generate an inverted-U. Two additional predictions of the model-that the average technological distance between leaders and followers increases with competition, and that the inverted-U is steeper when industries are more neck-and-neck-are both supported by the data.
Takeaway
This paper has investigated the relationship between product market competition and innovation. A Schumpeterian growth model was developed in which firms innovate ‘step-by-step’, and where both technological leaders and their followers engage in R&D activities. In this model, competition may increase the incremental profit from innovating; on the other hand, competition may also reduce innovation incentives for laggards. This model generates four main predictions which we test empirically. First, the relationship between product market competition (PMC) and innovation is an inverted U-shape: the escape-competition effect dominates for low initial levels of competition, whereas the Schumpeterian effect dominates at higher levels of competition. Second, the equilibrium degree of technological ‘neck-and-neckness’ among firms should decrease with PMC.1
U relationship between PMC and innovation in that industry. Fourth, firms may innovate more if subject to higher debt-pressure, especially at lower levels of PMC. Our empirical results confirm the existence of an inverted U-shaped relationship between product market competition and innovations, which in turn indicates that some kind of an escape-competition effect should dominate at lower levels of PMC as measured by the Lerner index, whereas the Schumpeterian effect pointed out in earlier endogenous growth models and before that in the IO literature, should dominate at high initial levels of PMC. Our results also indicate a similar inverted U-shaped relationship at the industry level, and that it tends to be steeper for firms that are in more neck-and-neck industries. Finally, we find that firms facing a higher threat of bankruptcy are subject to a stronger escape-competition effect and innovate more on average, especially at lower levels of competition. An important extension would be to introduce entry and entry threat as alternative (or complementary) measures of competition. This again would be done using an extension of the above model with entry and exit in any industry. Preliminary simulations performed on this extended model suggest an inverted U-shaped relationship between potential entry and the innovation rate.
Notes
- Philippe Aghion et al., Competition and Innovation: An Inverted-U Relationship, 120 Q.J. Econ. 701 (2005), https://doi.org/10.1162/0033553053970214. Takeaway source: 7. Conclusions, reproduced from a free author or repository copy (working-paper version). ↩
Cite as: Philippe Aghion et al., Competition and Innovation: An Inverted-U Relationship, 120 Q.J. Econ. 701 (2005), https://doi.org/10.1162/0033553053970214
Header image: Chromatogrphy lab.jpg by NeoLyo89, CC BY 4.0, via Wikimedia Commons.
