Wherever public schools are financed locally and enrollment is tied to residence status, cities control migrant children's access to public education. China's hukou system makes this margin measurable. We study it in a spatial, dynastic equilibrium model in which families choose where to work, whether to bring their children, and how much to invest in them. Access enters as the institution works: a brought migrant child is enrolled by lottery, drawn only after the family has moved. The lottery matters: a certainty-equivalent spending discount cannot produce the excluded migrant children observed in the data, and it understates the family-arrangement response by a third. Equalizing access reunifies families even as migration rises, raises migrant children's human capital, and improves intergenerational mobility. At the pace of budget adjustment we estimate from three decades of city data, the reform redistributes across generations rather than across lineages. Natives of restrictive cities pay through a temporary budget dip, but every birth lineage gains in the long run; at the lower end of the estimated fiscal elasticity, the most protected lineage roughly breaks even. Restriction itself has few beneficiaries: opening the megacities delivers the largest national gain of any single-tier access reform, without hurting the megacity-born.

Quantitative ex-ante urban-policy evaluation has historically relied on closed-form spatial-equilibrium and low-dimensional discrete-choice models that constrain analysis to homogeneous agents and fixed-form parametric utilities, and thereby to smooth proportional responses around the calibrated baseline. Integrating large language models into multi-agent simulation alleviates these constraints, but existing LLM-driven platforms lack the market-clearing mechanism and externally calibrated parameters that quantitative policy evaluation requires. We present LUSim (LLM-driven Urban Simulator), a multi-agent simulator in which heterogeneous households choose residence and workplace through a large language model and clear housing and labour markets inside an externally calibrated general-equilibrium city, with structural decision rules as baselines under identical conditions. We simulate a new rapid-transit line on a 96-zone Berlin instance calibrated from Ahlfeldt et al. (2015). The structural engines, which can only reweight the calibrated city in proportion, concentrate their response at the new stations, whereas with LLM-driven households approximately eight times as many relocate and the gains shift to peripheral districts. To test which engine predicts real urban change, we initialise every engine on the divided Berlin of 1986, let it predict the reunified city, and score, by criteria fixed in advance, its prediction of which districts gain or lose residents, jobs, prices, and wages against the changes observed by 2006. The simulation with LLM-driven households predicts these changes better than every structural baseline, also with district names withheld so that it cannot rely on memorised history, and the contrast persists across a second rail corridor, a zoning policy that expands floor space, and a second model family. We interpret these results as evidence that LLM-driven decision rules, disciplined by market clearing in a calibrated environment, predict urban change beyond the reach of structural models.