This project examines how marital agreements shape economic security and gender equality during marriage and at divorce. In Sweden, the legal default is equal division of marital property, a rule designed to protect the economically weaker spouse and to ensure against income losses related to specialization in unpaid care work. However, couples can opt out of this default by signing a marital agreement that establishes partial or full separation of property. Around one-fifth of married couples do so, yet we know remarkably little about who enters into such agreements, what they stipulate, and how different contractual choices influence economic trajectories over the life course. The Swedish public registry of prenuptial agreements dates back to 1921, with approximately 11,000 new agreements registered annually in recent years. This project constructs a longitudinal research database of prenuptial agreements from 1921 to 2024 by transforming scanned archival records into structured, analyzable data. The documents, currently available as scanned TIFF images will be processed using a reproducible document-understanding pipeline that combines optical character recognition (OCR), document layout analysis, and structured data extraction to produce document-level representations. Because the material contains sensitive personal information, all processing will be conducted within the Bianca secure computing environment for sensitive data. Building on the structured corpus, transformer-based language models will be used to analyze the semantic content of agreements, including the classification of asset arrangements, the identification of contractual provisions, and the analysis of linguistic and thematic change across cohorts and historical periods. The resulting dataset will be integrated within Statistics Sweden’s (SCB) secure research environment, enabling linkage to demographic and socioeconomic registers under strict data protection controls. The project builds a unique population-wide database covering all marital agreements registered in Sweden since 1970. These contracts are digitized and linked to rich administrative register data on income, employment, capital income, business ownership, demographic characteristics, and family structure. This linkage enables following individuals and couples before and after contracting, throughout marriage, and in the event of divorce. The research is organized into three integrated work packages. First, we document long-run trends in marital agreements and identify which individual, couple, and extended-family characteristics predict selection into contracting. Second, we analyze the content of agreements and examine selection into different degrees of property separation. Third, we exploit variation in contractual design—comparing full to partial separation—to estimate how stronger individualization of property rights affects labor supply, within-couple income gaps, and economic vulnerability at divorce, using longitudinal data and matching methods to address selection. By making marital agreements empirically observable at scale, the project generates new evidence on how private contracting within marriage interacts with legal institutions to shape economic inequality, gender equality, and financial security, with direct relevance for research and public policy.