The contingent cooperative: state infrastructure and the juridical-form composition of new organisations in Misiones, Argentina, 1901-2025 (data and code)

Data and analysis code for a population-level study of the formal organisational landscape of the Argentine province of Misiones, 1901-2025. The deposit contains the Misiones subset of the Registro Nacional de Sociedades (14,278 organisations; 14,169 dated), Argentina's annual real-GDP growth series 1961-2025 (World Bank, World Development Indicators, series NY.GDP.MKTP.KD.ZG), and the canonical model-free analysis: Shannon diversity with bootstrap confidence intervals, juridical-form composition with exact Poisson intervals, a functional disaggregation of cooperatives, and a permutation test of sub-departmental concentration. The central finding is a contingency asymmetry: self-sustaining commercial forms diffuse secularly across state projects of opposed political orientation, whilst the cooperative, whose creation depends on a continuous social-economy apparatus, surges when that apparatus is sustained and returns to its closed-channel share when it is withdrawn. That share occupies 4.1-4.8 per cent of new registrations in every period in which the channel through which cooperatives are constituted was closed, and 14.8-25.5 per cent in every period in which it was open, without overlap - whilst the same periods overlap in mean output growth, so the macro-economic cycle does not separate what the channel separates. An earlier correspondence-analysis and sequence pipeline was abandoned because entering the period of creation as an active variable makes the principal axis an axis of time and manufactures a spurious terminal convergence. That pipeline and the diagnostics establishing the defect are retained in archive/ for transparency; no result in the paper is computed from them. The ARCA fiscal-status enrichment (enriquecido_arca_v2.parquet, tab_fiscal_status.csv) is likewise retained but supports no claim in the paper: the 2024 regulatory action it would have to control for suspends an entity's authorisation to operate, which is a different register from the tax record the measure reads. Code is released under the MIT licence; data and figures under CC-BY-4.0.