Quantitative water balance assessments of traditional irrigation systems remain scarce, particularly for systems experiencing chronic water deficits driven by institutional rather than climatic factors. This study applied the F.J. Mock rainfall-runoff model to Subak Balangan, which has experienced a 20-year-long continuous water shortage due to an inter-district allocation dispute. Using 10-year climate data (2015-2024) and independent field validation measurements (March-July 2025, 10 half-monthly periods), the model achieved moderate-to-good performance ( NSE = 0.521 R ^ 2 = 0.833 RMSE = 0.01 m³/s). Water availability at 80% reliability ranged from 0.001 to 0.012 m³/s across 24 half-monthly periods. A comparative water balance analysis of two crop rotations revealed that the existing flower-cassava rotation experienced deficits in 75% of periods (maximum 0.017 m³/s), while a hypothetical paddy-paddy-maize rotation would produce deficits in 91.7% of periods, with 2.96-fold higher peak demand (maximum 0.074 m³/s). These results quantify the magnitude of water stress under existing institutional constraints and demonstrate the limitations of paddy-based crop intensification as a water management strategy in this specific context. Beyond empirical findings, the study advances two theoretical contributions: Institutional Hydro-Decoupling (IHD), extending Ostrom's common-pool resource theory to conceptualise conditions where governance failure structurally severs the link between catchment hydrology and effective irrigation supply; and Hydro-Adaptive Cropping Moderation (HACM), extending autonomous adaptation theory to identify farmer-led rotation adjustment as an endogenous demand-side mechanism that partially compensates for institutionally-suppressed water availability. The findings provide a baseline water balance reference for ongoing allocation dispute resolution and highlight the need for multi-Subak comparative studies to establish broader patterns.