The Dynamics of a Policy Outcome: Market Response and Bureaucratic Enforcement of a Policy Change

GAICIL Insights

A statute is a request addressed to a market.¹ Whether the market complies is another matter. The Dynamics of a Policy Outcome: Market Response and Bureaucratic Enforcement of a Policy Change, by Steven Callander, Dana Foarta and Takuo Sugaya, starts from a deceptively simple observation: policy outcomes are determined by the actions of private citizens. A policy succeeds or fails through how it shapes behavior and how that behavior shapes the future course of policy.

For AI lawmakers, the sting is this: before adding another article to a statute, fund the agency that will enforce the first ones. On this model, the enforcer's credibility may decide whether the rules count for anything.

To capture this, the authors build a model joining the political and nonpolitical domains, focused on competition policy and market regulation. The outcome of a policy change develops over time as firms respond in the market and interact with bureaucratic enforcement. A critical threshold in market structure determines whether a policy succeeds or fails, and the design of political institutions affects where that threshold lies.

The threshold balances two forces. Politics is path dependent; markets are self-correcting. The threshold marks when political forces dominate market forces and therefore when a policy change has lasting effects on society. Below it, market forces may undo a reform. Above it, the reform takes hold and shapes outcomes durably. Reform is a tug of war with a line on the ground.

The logic applies everywhere. Outcomes in every area of politics are only partially within legislators' control. Legislators can at most set guidelines and provide incentives; private actors determine how behavior, and thus policy outcomes, form. Imperfect control fundamentally shapes policymaking and the practice of politics. The legislator proposes; the market disposes.

Bring the result forward to the present. AI governance is the perfect test case. AI policy is being written in many jurisdictions, and its effects will depend on how developers, deployers and users respond. Rules on AI safety, transparency or competition will be interpreted, implemented and sometimes circumvented by market participants, and enforced by bureaucracies with limited resources. The success of AI regulation cannot be read from the text of statutes. Word counts of AI laws measure ambition, and the threshold decides results.

The threshold result is a forecasting tool. Where market structure and dynamics favor a policy's aims, reforms can take hold. Where powerful forces push against them, reforms may erode over time. Policymakers should assess market structure and likely firm responses before expecting regulations to deliver. An AI rule written for a market that will route around it is a press release with legal formatting.

Speed raises the stakes. AI markets evolve quickly, and enforcement agencies may struggle to keep pace. The self-correcting nature of markets can be a strength, letting competition discipline behavior. It can also render poorly designed rules quickly irrelevant. Recognizing these dynamics helps policymakers design rules that work with market forces.

Institutions move the line. Their design affects the threshold that determines success. Stable, credible enforcement may lower it, letting reforms take hold more easily. Unstable or captured institutions may raise it. For AI governance, building credible, competent enforcement institutions may matter as much as the content of the rules. An excellent rule with a weak enforcer may lose to a modest rule with a strong one.

Humility follows. Legislators cannot fully control outcomes, and attempts to micromanage complex markets may fail. Policies setting clear guidelines and incentives, leaving room for private actors to innovate within them, may outperform detailed mandates that markets will reshape anyway.

Pulled together, policy is a dynamic process. Outcomes emerge from the interplay of rules, markets and enforcement over time. This paper provides a rigorous framework for analyzing that interplay in competition policy and beyond. For AI governance, where stakes are high and markets move fast, the real test of a policy is how it shapes behavior.