Digital capacity and the conditional effectiveness of green bonds in global climate finance markets

Green bond markets have expanded rapidly, but their contribution to emission reduction remains unsettled. This paper asks whether green bond effectiveness is conditional on digital capacity, measured as the information infrastructure that carries disclosure, traceability, verification and post-issuance scrutiny. The analysis uses an annual panel of 76 countries over 2007–2021 and combines two-way fixed effects, threshold fixed effects, panel smooth-transition regression and dynamic common-correlated-effects estimation. Green bond issuance shows no stable unconditional association with CO 2 emissions per capita. Its marginal effect is positive or statistically indistinguishable from zero in low digitalisation environments, and becomes negative as countries enter a higher range of digital capacity. The continuous specification places the zero-crossing near the 59th percentile of the digitalisation distribution, while the threshold model identifies a regime split around the 70th percentile. These estimates are identified on the margin of actively issuing economies: 219 of 1072 country-year observations record positive issuance, and 119 of them, from 28 countries, fall inside the high-capacity regime. Financial deepening follows a similar but weaker pattern: bank assets remain positively associated with emissions, though this association declines as digital capacity improves. Digitalisation and institutional quality are difficult to separate because they co-vary strongly within countries. Their common capacity component better explains when green bonds become associated with lower emissions. The pattern is consistent with an enforcement reading, in which the environmental value of climate finance depends on the digital and institutional architecture that makes claims checkable.