Flow Variability and Data Sharing in Transboundary River Treaties

Treaties on transboundary river sharing often include provisions for sharing hydrologic data, yet such sharing is often limited where hydrologic variability is high. We study this puzzle in a model where treaty obligations must be conditioned on a noisy but verifiable public signal of inflows, and improving its accuracy is costly. We find that greater variability reduces the value of signal accuracy under concave benefits. When investments in accuracy are part of a negotiated treaty, the effect of variability on negotiated signal accuracy can be non-monotonic. Using treaty-text codings matched to basin-level variability measures, we provide descriptive evidence that basins with more variable river flow tend to feature weaker data-sharing provisions. The findings highlight a constraint on treaty design: when treaties can condition only on noisy but verifiable information, higher variability need not lead to stronger data sharing.