Why do good policies so often fail at the point of implementation?

The usual explanation is that officials are incompetent or corrupt. Both can be true, but neither is the deep problem. A well-designed policy fails because of the structure of rules and the knowledge that the people carrying it out work within.

Consider what happened during the COVID-19 lockdown in 2020. The government sensibly permitted inter-state trucking for essential goods to keep supply chains running. But the trucks did not move. It took officials in Delhi weeks to work out why, and the answer was obvious to anyone actually on the highways: the dhabas and roadside repair shops had been shut along with everything else. Without anywhere to eat, rest, or fix a breakdown, no driver can cover 600 kilometres through the night. The policy allowed trucking. It just failed to account for the basic conditions that make trucking possible.

The economist Friedrich Hayek called this the knowledge problem: no central authority can gather and process all the dispersed, local, and tacit knowledge that determines what works in practice. The officials who wrote the lockdown rules did not drive trucks and did not consult anyone who did. They assumed they understood the system, and the supply chain broke before they found out they did not. 

This is not a problem confined to emergencies. Policies are designed in Delhi or the state capitals and then implemented across thousands of districts, blocks, and villages, each with its own conditions. A scheme built around a generalised idea of the rural household meets realities in Vidarbha, Bundelkhand, and coastal Odisha that its designers never imagined. That gap is not a failure of effort. It is a limit on what centralised design can know.

Incentives then widen the gap. Officials who implement a programme are judged on compliance: were the procedures followed, the reports filed, the targets met on paper? They are rarely judged on whether the programme worked. The Swachh Bharat Mission shows both failures at once. The government declared that toilets had been built for all rural households and open defecation had been eliminated. But a Comptroller and Auditor General audit found that in 41 of 120 Gram Panchayats it tested in Gujarat, the toilets stood unused because no water connection had ever been extended to them. The knowledge problem recurs here: a toilet without water is a structure, not a sanitation system, and the local condition that made it useless was invisible from the centre. The incentive problem completes the picture. The official’s task was to build the physical structure, and it was completed. Whether anyone used the toilet was nobody’s responsibility, and the gap between toilets built and toilets used extended well beyond Gujarat.

So the pattern holds wherever the people who design and deliver a policy bear no personal cost when it fails. As long as an official is rewarded for following the procedure rather than for solving the problem, the distance between what a policy promises and what it delivers will remain. A policy is only ever as good as the incentives facing the people who carry it out.