Cheap Exit, Unmoved Marriage: Aggregate Neutrality and the Completed Elite Filter of American Marriage, 2000-2024

The completion of the no-fault divorce revolution - New York, 2010 - is conventionally indicted as the final blow to the American marital institution. We test this claim on 66,409,453 complete-case person records from IPUMS USA (2000-2024) using a three-pass streaming linear probability architecture with inference disciplined for the single-treated-cluster structure. The baseline estimate is positive and conventionally significant (+0.57 pp; t = 3.98, CR1 p = 2.2e-4) but does not survive valid inference: a restricted wild cluster bootstrap (Webb weights, 999 replications, null imposed) yields p = 0.471, and a synthetic control counterfactual produces a post-2010 gap of -0.41 pp indistinguishable from the in-space placebo distribution (exact permutation p = 0.667; post/pre RMSPE ratio 2.85, rank 34 of 51). We read this aggregate neutrality as completion, not absence, of structure: by 2000 the participation margin had closed to exit-cost shocks and the institution's action had migrated to the compositional margin (college +15.31 pp, t = 32.56; log income +2.06, t = 30.79). Volume II of The New Human Race project.