What happens when part of a city cuts minimum lot sizes? We study a 2014 reform where a council in Adelaide cut them from 350 sqm to as low as 150 sqm, until the policy was partly reversed in 2019. The reform increased building approvals by 67 per cent, driven entirely by townhouses and rowhouses; the reversal returned the treatment effect to insignificant levels, providing dose-response evidence that lot-size regulation was a binding constraint. Established local firms already producing medium-density housing primarily built the new homes, with no evidence of capacity diversion. Movers into new housing were predominantly households in their prime family-formation years, upgrading out of cheaper rental housing. Using linked Census microdata, we trace the movement chains they set off: each mover vacates a cheaper dwelling, occupied by a household further down the income distribution. Chains also track the lifecycle: occupants further along the chain are younger and more often renting, with age composition explaining part of the income gradient. Concerns about amenity costs drove the reform’s reversal; across the outcomes we can measure we find little deterioration.