On the Welfare Effects and Political Economy of Competition-Enhancing Policies

GAICIL Insights

The standard cost-benefit spreadsheet may be systematically undervaluing competition.¹ On the Welfare Effects and Political Economy of Competition-Enhancing Policies, by Philippe Aghion and Mark Schankerman, analyzes the welfare effects of policies intensifying product market competition and the channels through which they work, using a circular model of horizontal product differentiation with asymmetric costs.

Three channels are in play. Increased price competition affects market selection among incumbents, cost-reduction effort and entry of new firms. Competition can drive less efficient firms out, push incumbents to cut costs and encourage efficient entrants. Each channel contributes to the welfare effects of competition-enhancing policies. Competition cleans house, sharpens the survivors and opens the door.

Simulations show selection effects can be large, and traditional cost-benefit analysis is unlikely to capture them. Conventional approaches may therefore underestimate the benefits of policies and infrastructure intensifying competition. Magnitudes depend on the economy's characteristics, or initial conditions. The biggest gains are the ones the standard toolkit does not see.

Political economy is the twist. Cost heterogeneity generates endogenous demand for policies that enhance or retard competition. Efficient firms may favor competition; less efficient firms may oppose it. The authors show the possibility of a low-competition political economy trap, where an economy stays stuck with low competition because those who would lose from more competition resist reform. The losers from reform are often the best organized.

The application to AI is direct. Artificial intelligence can intensify competition by lowering costs, enabling new entrants and increasing price transparency. The welfare benefits may be large and missed by traditional analysis. Evaluating AI's market effects requires attention to selection, cost reduction and entry. A review of AI's economic impact that counts only price changes is counting the smallest channel.

Selection is the hidden prize. AI may let more efficient firms expand and less efficient firms exit, raising average productivity. These reallocation effects can generate substantial gains conventional measures overlook. Recognizing them strengthens the case for letting competition operate.

The trap is the hidden danger. Firms threatened by AI-driven competition may seek policies limiting it, such as restrictive regulations or entry barriers. If such firms are influential, the economy may become trapped in low competition, forgoing AI's benefits. Some calls for AI regulation will come from those who fear AI-enabled rivals, and the model explains why. Understanding the dynamics helps policymakers anticipate and address resistance.

Infrastructure deserves note. Policies and infrastructure intensifying competition include investments reducing barriers between markets. Digital infrastructure enabling AI-driven services to reach consumers can likewise intensify competition and generate welfare gains, possibly with larger returns than traditional analysis suggests. Connectivity is competition policy in concrete form.

Initial conditions matter. Effects of competition-enhancing policies vary across economies. Policymakers should assess their economy's characteristics, recognizing gains may be particularly large where competition is initially weak. The economies that most need competition may have the most to gain from AI.

The modeling contributes to competition analysis. Incorporating asymmetric costs, selection and entry captures channels simpler models miss, and similar frameworks can inform evaluation of AI markets and sectors AI transforms.

Policy follows. Competition-enhancing policies can generate large welfare gains, and political economy forces may resist them. Accounting for selection effects and anticipating opposition from less efficient incumbents helps reforms succeed.

Pulled together, intensified competition brings substantial welfare gains through selection, cost reduction and entry, gains traditional analysis may miss. Political economy dynamics can trap economies in low competition. As AI opens new opportunities for competition, recognizing these gains and addressing resistance will help economies capture technological progress's full benefits.

Something for regulators hearing incumbents' warnings about AI to sit with: ask who would lose market share if AI-enabled rivals were free to compete. In the model's trap, the loudest voices for restraint are the firms that competition would select out.