Should cleanliness drives be incentivized to encourage people in India?

Incentives pay for what can be measured, and the measure is easy to game

Question: Should cleanliness drives be incentivized to encourage people in India?
Position‹2 of 2: No, cleanliness drives should not be incentivized to encourage people in India
Argument‹2 of 2: Incentives pay for what can be measured, and the measure is easy to game

The argument

The practical objection is that cleanliness cannot be paid for directly, because it cannot be measured directly, and that everything which follows from this undermines the scheme. Critics point out that an incentive must attach to something countable. In practice that means attendance at a drive, bins installed, photographs submitted, households registered, or weight of waste collected. None of these is cleanliness. Each is a proxy chosen because it can be recorded, and the gap between the proxy and the outcome is where the scheme fails. Proxies are manipulable, and an incentive gives people a reason to manipulate them. Waste can be moved rather than removed, and the same waste can be collected more than once. Attendance can be recorded without work being done. Photographs can be taken of a cleared frontage while the lane behind it is untouched, since the record only requires one image. Critics stress that none of this depends on participants being dishonest, only on their responding to what is actually being rewarded, which is the production of the record rather than the condition of the place. The targeting problem compounds it. Payments flow to wherever measurement is cheapest and participation is easiest to organise, which tends to be the areas already in better condition and already better served. The places with the worst sanitation are the hardest to register, audit and photograph, so a scheme that pays for verified activity reliably spends least where need is greatest. On this account incentives do not merely fail to buy cleanliness; they divert attention and money toward the appearance of it, and that appearance then becomes the evidence on which the scheme is judged a success.

Premises

[P1]Cleanliness cannot be measured directly, so an incentive must attach to a countable proxy such as attendance, bins installed, photographs or weight collected. [P2] Proxies are manipulable, and rewarding them gives participants reason to produce the record rather than the outcome, without requiring any dishonesty. [P3] Measurement and verification are cheapest in areas that are already better served, and hardest in the places with the worst sanitation. [P4] Payments therefore concentrate where need is least, and the resulting records become the evidence by which the scheme is judged. [C] Cleanliness drives should not be incentivised, because the payment can only reach proxies that are gamed and mistargeted, and the appearance of success obscures the failure.

Counter-arguments

Proponents reply that this is an argument about scheme design rather than about incentives in principle. Every public programme pays against imperfect measures, and the response is auditing, randomised verification, rewards tied to independently assessed outcomes rather than to self-reported activity, and payment at the level of an area rather than an individual, which removes the gain from moving waste next door. They also argue the targeting objection points at a fixable allocation rule: weighting payments toward the worst-served areas is available to any designer who wants it, and refusing incentives altogether leaves those same areas with nothing.

Rejecting the premises

[Rejecting P2] Gaming is a design failure rather than a property of incentives, and auditing, randomised verification and rewards tied to independently assessed conditions rather than self-reported activity all reduce it. [Rejecting P3] Paying at the level of an area rather than the individual removes the gain from moving waste next door, which disposes of the clearest example the argument relies on. [Rejecting P4] Concentration of payments on better-served areas follows from an allocation rule, not from incentives themselves, and weighting payments toward the worst-served areas is available to any designer who wants it.