The two previous lectures established that verifiable procedures are technically achievable without sacrificing ballot secrecy. This lecture opens with the question that follows immediately: if that is true, why are such systems still the exception rather than the norm? The convenient answer — that the technology simply isn't mature enough — is not false, the lecture argues, but it systematically overlooks a more important one: institutional change produces losers as well as winners, and wherever it produces identifiable losers, resistance is a predictable, rational response, not an anomaly. The lecture is explicit that this frame concerns institutions and incentive structures as a class of phenomena, not the personal character of any individual who happens to resist reform.

The first substantive block supplies the economic vocabulary: rent-seeking — the deployment of resources not to create value but to capture existing value by shaping rules in one's own favor. Opacity, the lecture argues, is very often not an accidental byproduct of imperfect design but a valuable asset in its own right, since it lowers both the cost of acquiring rent and the risk of losing it. This yields a genuinely testable hypothesis, central to the rest of the lecture: resistance to transparency should concentrate specifically where extractable rent is greatest, not spread evenly as ordinary organizational inertia would predict — a distinction that is, in principle, empirically observable by looking at where resistance clusters.

The core block introduces principal–agent theory to explain the concrete mechanism, including cases involving no wrongdoing at all. Because an agent always knows more about their own routine work than the principal who delegated authority to them, monitoring — audits, reporting, review — is never free: it consumes time, diverts attention, and carries a real risk of unfair suspicion. From this follows the lecture's central methodological distinction: technical resistance, grounded in genuine implementation costs, versus political resistance, which defends a position that opacity itself makes possible. The two sound alike and manifest identically — delays, hedged qualifications, calls for further study — so telling them apart requires analyzing where objections cluster relative to where rent is extracted, not taking stated motives at face value.

A deliberate counterpoint follows, built around Jonathan Fox's influential critique of the assumption that transparency automatically produces accountability. Fox separates two independent axes: transparency can be "clear" or "opaque" — real access to traceable information, or a formal release of raw, unprocessed data that structurally resists analysis — while accountability can be "soft", a bare requirement to explain, or "hard", carrying real consequences. The practical payoff: institutions frequently welcome reforms labeled "transparency" precisely because opaque transparency is cheaper and safer than clear transparency, producing the appearance of reform without any of the accountability that would actually threaten anyone's position — illustrated through the historically distant, politically neutral cases of freedom-of-information reform in Mexico and, later and more sweepingly, India.

The lecture then grounds all of this in a historical illustration chosen deliberately at a safe distance from the present: turn-of-the-twentieth-century patronage systems, in which public offices and contracts were distributed on personal loyalty rather than verifiable criteria — opacity here was not an incidental flaw but the system's structural foundation, since explicit, equally applied criteria would eliminate the very rent that discretion made possible. The civil service reforms that introduced competitive examinations met prolonged, organized resistance whose intensity, historians document, tracked not the technical difficulty of the reform but the size of the rent it eliminated at each particular point — directly confirming the lecture's opening thesis.

The lecture closes by drawing these threads into one line of argument and opening the question that carries into the rest of the series: if both outright resistance and the mere imitation of reform are predictable, rational responses to a shift in incentives, what would a verifiable procedure need to resist being hollowed out over time — and could the role of auditor be filled not only by a human, but by an artificial intelligence?