Competition and productivity growth in South Africa

GAICIL Insights

Comfortable firms make poor innovators, and South African manufacturing was very comfortable.¹ Competition and productivity growth in South Africa, by Philippe Aghion, Matias Braun and Johannes Fedderke, shows mark-ups significantly higher in South African manufacturing industries than in corresponding industries worldwide, then tests what this weak product market competition did to productivity growth.

The answer is stark. High mark-ups have a large negative impact on productivity growth in South African manufacturing. Reducing mark-ups, meaning more product market competition, should have large positive effects on productivity growth. The findings hold across three data sources, two alternative measures of productivity growth and three distinct mark-up measures. Controlling for potential endogeneity of regressors does not eliminate them. Profits that look like success on a balance sheet can be a symptom of national underperformance.

Consistency is the persuasive force. Using three alternative panel data sets, the authors assess competition and then estimate its growth effect. Each time, mark-ups are high relative to international benchmarks, and higher mark-ups go with lower productivity growth. Competition emerges as a significant lever for raising productivity. Three independent views converge on one picture.

Extensions are proposed: better instruments for product market competition, analysis of entry measures and their relation to growth, and a closer look at the link between trade liberalization, competitive pressure and productivity growth. These would deepen understanding of competition and growth in emerging economies.

Now the AI angle. Artificial intelligence adoption depends on the same pressure. AI can raise productivity, and the incentive to adopt depends on competition. Firms enjoying high mark-ups in protected markets may have weaker incentives to invest in new technologies. More competition could spur AI adoption and its productivity gains. A firm that can raise prices has little reason to adopt AI.

The logic applies to AI markets themselves. Where AI providers enjoy high mark-ups, diffusion to other sectors may slow because of high prices. Competitive AI markets, with lower mark-ups, can speed diffusion and amplify productivity benefits across the economy. Expensive AI is slow AI.

Measurement is robust. Three distinct mark-up measures guard against results reflecting a single flawed indicator.

The international benchmark is a practical tool. Comparing domestic mark-ups with those in corresponding industries worldwide reveals where competition is weak. The same comparison could be applied to digital and AI-related markets to identify where competition policy might yield large gains. A country can learn a lot by asking why its firms earn more than their foreign peers.

Trade liberalization is a natural channel. Openness increases competitive pressure, which can reduce mark-ups and raise productivity. For emerging economies, trade openness in technology goods and services may bring AI and its benefits to domestic firms while spurring them to improve. Imports can carry both the technology and the pressure to use it.

For emerging-economy policymakers, competition policy is growth policy. Reducing entry barriers, opening markets to trade and enforcing competition law can lower mark-ups and raise productivity growth. As AI offers new productivity opportunities, a competitive environment helps ensure firms seize them.

The robustness is a methodological strength. Testing across multiple data sources and measures, and addressing endogeneity, gives credible evidence on a question central to development policy, valuable for South Africa and economies with similar characteristics.

Distilled to one claim: weak competition reduces productivity growth, and more competition can raise it substantially. In South African manufacturing, high mark-ups had a large negative impact on productivity growth. As AI becomes a worldwide source of productivity gains, competitive markets will help economies, especially emerging ones, capture them by spurring adoption and ensuring new technology's benefits reach consumers and firms.

Something for emerging economies drafting AI strategies to sit with: start with a mark-up audit. If domestic firms earn far more than global peers, the binding constraint on AI adoption may be the absence of anyone forcing them to adopt.