Regulations interact, and the interaction is usually invisible in the individual impact assessment. Take three rules that were each assessed alone. A capital rule makes a bank hold more equity against small-business lending. A disclosure rule makes securitisation expensive. A priority-sector target requires the small-business lending anyway. Together they produce an outcome none of the three intended, because the lending migrates to a lightly regulated non-bank that funds itself from the same banks through a channel the capital rule prices differently.
Each rule was assessed alone. The system was not.
What is regulatory capture, and how does it arise?
Regulatory capture is the theory that a regulator comes, over time, to act in the interest of the industry it regulates rather than the public it was created to protect. George Stigler set out the economic version of the argument in 1971, and the mechanism is not corruption.
It is arithmetic.
The three preconditions let you diagnose capture from structure rather than from suspicion. Asymmetric stakes: the industry has a large concentrated per-firm interest in the rule while the public has a small diffuse per-person one. Twenty firms splitting a ₹2,000 crore benefit each have ₹100 crore of reason to act; a hundred million citizens splitting a ₹2,000 crore cost each have ₹200 of reason, which will not pay for a lawyer. Asymmetric information: the regulator's technical understanding comes overwhelmingly from the industry, because that is where the engineers and the data are. Asymmetric attention: the industry is present at every consultation and every comment period, and the public is present at none.
Two documented patterns, stated as public record rather than as accusation. The first is the revolving door, where staff move between the regulator and the regulated in both directions; it is measurable from public appointment records and is not by itself proof of anything, since it also transfers genuine expertise. The second is the pattern found repeatedly in post-crisis official inquiries, where a supervisor with formal power declines to use it because the supervised firm's own model, reviewed by the supervisor's own staff, said the risk was small.
Neither pattern requires anybody to be dishonest, which is exactly what makes capture hard to legislate against.
Advertisement
How do you guard against regulatory capture?
The countermeasures follow from the preconditions. Sunset clauses force a rule to be re-justified rather than to persist by inertia, and cooling-off periods slow the revolving door. Mandatory publication of every meeting with a regulated party attacks asymmetric attention, and independent technical capacity funded outside the industry levy attacks asymmetric information. Read a regulatory regime looking for those four and note which are absent, because a regime with none of them drifts toward the industry regardless of who runs it.