Is a UBI pilot actually useful evidence? The methodological problems

There have been now several UBI pilots — Finland, Stockton CA, Kenya GiveDirectly, and others. The optimism about what we can learn from these is, I think, overblown. **Sample size and selection**: Most pilots are small (hundreds to low thousands) and volunteer/lottery-based. The macro effects of UBI (inflation, labour supply at scale) simply cannot be measured in a pilot. **Duration**: A 2-year pilot doesn't tell you about long-run behaviour change. People know it ends; they make decisions accordingly. **General equilibrium**: If everyone got UBI simultaneously, the economy adjusts in ways a single-city pilot cannot capture. This doesn't mean pilots are worthless. They produce useful data on individual welfare outcomes, mental health, and some spending patterns. But treating them as evidence about 'whether UBI works at scale' is a category error. For/against UBI: the arguments are fundamentally about values and macro theory, not about what the Stockton results showed.