When the State Enters the Bedroom: Mandatory Arrest and the Boundaries of the Marital Contract, 1990-2010

Between 1985 and 2002, twenty-three U.S. states stripped police officers of arrest discretion in domestic-violence incidents. We treat each staggered adoption of a mandatory-arrest statute as a shock to the implicit enforceability of the marital contract and estimate effects on the stock and dissolution margins of intact marriages using 36,867,710 person records from IPUMS USA (1990, 2000-2010), sliced from two existing verified panels. In a micro-level linear probability model absorbing state and year fixed effects and state-specific linear trends (51 clusters, CR1, test df = 50), mandatory arrest leaves the married stock unchanged short-run (+0.12 pp, t = 0.35) and long-run (+0.40 pp, t = 0.99), and leaves separated (-0.20 pp) and divorced (-0.18 pp) margins unmoved. A state-by-year-by-gender triple-difference finds no within-household incidence (-0.26 pp, t = -1.17); Callaway-Sant'Anna event estimates confirm the aggregate null (ATT = +1.1 per 1,000, bootstrap SE = 4.2). Read alongside Iyengar (2009), the comprehensive null bounds the channel: bedroom interdiction repriced violence inside incumbent marriages, not the marriages themselves. Volume III of The New Human Race project - the boundary-condition study.