Green Finance Works Best Where Governance Is Strong: New Evidence From Vietnam’s Provinces

Vietnam has spent the past decade pouring credit into renewable energy, clean agriculture, and low-carbon industry, but a new study suggests that the money alone does not decide whether a province actually transforms its technology base. Researchers have built a province-by-province indicator of environmental technology transition for all 63 of Vietnam’s provincial-level units and tracked it from 2014 to 2024, finding that green finance delivers measurable clean-technology gains only when local institutions are strong enough to convert loans into real innovation. In provinces with weak governance, the same infusion of green credit produces almost nothing.

The study, published in Environmental and Sustainability Indicators, addresses a stubborn measurement problem. Most research on green finance relies on single outcomes, such as counts of green patents or installed renewable capacity, which capture only one slice of a complex transition. The Vietnamese team instead constructed a composite Environmental Technology Transition index using principal component analysis, blending four theoretically grounded components: granted patents in the IPC Y02 climate-mitigation classes, the share of installed renewable capacity in each province’s power mix, environmental fixed-asset investment as a proportion of total investment, and energy-efficiency performance measured as economic output per unit of final energy use. The first principal component explains 54.5 percent of the variation across the panel, with component loadings between 0.61 and 0.78, indicating that no single metric dominates the composite.

The data reveal striking heterogeneity. Renewable-energy penetration ranges from zero to 68.9 percent across provinces, and environmental investment intensity spans from 0.3 to 22.6 percent of fixed-asset investment. That variation is precisely what makes Vietnam, with its 58 provinces and five centrally administered municipalities spanning eight socio-economic regions, an ideal laboratory for asking whether finance, geography, and governance jointly drive clean-technology upgrading. The study period is anchored by two policy milestones: the 2012 National Green Growth Strategy, which established national green-credit targets, and the country’s 2021 COP26 pledge to reach net-zero emissions by 2050.

Measuring the effect of green credit is notoriously difficult because banks do not allocate loans randomly. Lenders may simply channel green credit toward provinces that were already environmentally progressive, creating a correlation that looks causal but is not. The researchers confronted this endogeneity with two unusual instruments. The first draws on colonial history: the log count of French-chartered banking branches per 10,000 residents during the Indochina period from 1900 to 1954, reconstructed from French colonial archives and the Banque de l’Indochine’s provincial registries, interacted with a post-2015 policy indicator. The second exploits climate exposure, using annual flood-event frequency from disaster records as a source of financing demand that is plausibly unrelated to clean-technology outcomes once fixed effects and province-specific trends are controlled.

The headline result is a structural coefficient of 0.334 from a spatial two-stage least squares estimator: a one-standard-deviation increase in green finance intensity is associated with a 0.334-standard-deviation rise in the environmental technology transition index, conditional on controls, province and year effects, and province-specific linear trends. The instrument diagnostics are strong, with a Kleibergen-Paap first-stage F-statistic of 40.1, well above conventional weak-instrument thresholds, and an Anderson-Rubin confidence interval running from 0.212 to 0.461. A temporal placebo using pre-policy data from 2010 to 2013 finds no comparable association, and a spatial placebo that randomly permutes the geographic weight matrix across 1,000 permutations yields nothing, suggesting the pattern is specific to real geography and the post-2015 green-credit window.

Perhaps the most striking finding is spatial. Environmental technologies do not respect provincial borders. Using a spatial Durbin model with a row-standardised Queen contiguity matrix, the study finds significant positive spatial autocorrelation in both the transition index and green credit allocation, with Moran’s I statistics of 0.192 and 0.176 respectively. The indirect, cross-boundary impact of green finance is estimated at 0.189, meaning that a province’s neighbours’ green lending spills over into its own technology transition through knowledge flows, demonstration projects, supply-chain linkages, and skilled-labour mobility. The spatial feedback multiplier of 1.30 implies that outcomes reverberate through the network rather than stopping at each border. In practical terms, a province surrounded by green-finance-rich neighbours benefits even if its own banks stay cautious.

The transmission channels behind the aggregate effect are also decomposable. Bootstrap mediation analysis with 1,000 replications attributes 22.1 percent of the total effect to research and development investment and 15.7 percent to clean capital formation, leaving roughly 62 percent as the direct effect of relaxing financing constraints for cleaner capital goods. The R&D channel is far stronger in well-governed provinces, accounting for 28.4 percent of the effect there versus just 11.2 percent in weakly governed ones, consistent with the idea that directed technical change requires institutional infrastructure to translate financial flows into innovation output.

That institutional dependence is the study’s sharpest policy warning. Splitting the sample at the median of the Green PCI environmental governance score, the researchers estimate that high-institutional-quality provinces reap 0.421 standard deviations of transition gain per unit of green finance, while low-quality provinces see only 0.169, an approximately 2.5-fold differential. A Hansen threshold test further identifies a minimum green-credit intensity of roughly 0.287 standard deviations below which the transition mechanism is not fully activated. In weakly governed provinces, the authors argue, green credit risks being absorbed into low-additionality, compliance-only uses, a pattern the literature describes as institutional greenwashing, and the risk is most acute exactly where monitoring capacity is weakest. Substituting Rule-of-Law scores for the governance index produces a nearly identical pattern, suggesting the result is not an artefact of one survey.

The robustness battery is unusually thorough. The green-finance coefficient stays positive and significant across inverse-distance, road-network, and economic-proximity spatial weights, after excluding the two dominant metropolitan hubs of Hanoi and Ho Chi Minh City, after dropping the reconstructed 2014 to 2016 green-credit data, after excluding the Covid-disrupted years, after winsorising outliers, and in dynamic GMM and event-study designs built around Vietnam’s 2022 environmental protection law. Alternative outcomes tell the same story: patent-only measures, renewable-energy shares, and economic-only index variants all yield significant positive coefficients, while energy intensity and emission intensity both fall significantly as green finance rises.

For policymakers, the implications cut against simply expanding lending everywhere. The authors propose a tiered green taxonomy that distinguishes deep eco-innovation projects from incremental improvements and compliance-minimum spending, with concessional terms tied to verified innovation additionality and a project-level green credit registry linking loans to patenting, renewable adoption, and efficiency outcomes. Because spillovers are real, regional coordination of green-finance programmes is more defensible than isolated provincial efforts. And because the diagnosis matters, an indicator-based accountability system can distinguish provinces that need more credit from those that need better project screening and implementation capacity, and mobilise strong performers as regional diffusion hubs. The broader lesson travels beyond Vietnam: in emerging economies racing toward net zero, money is necessary but not sufficient, and the binding constraint on clean-technology transition may be the quality of the institutions that spend it.

Subject of Research: Green finance, spatial diffusion, and institutional absorptive capacity in provincial environmental technology transition in Vietnam

Article Title: A provincial environmental technology transition indicator for Vietnam: Green finance, spatial diffusion, and absorptive capacity

Article References: Tran, A. Q., Pham, M. T., Trinh, K. C., Pham, T. L., Pham, D. K., Pham, T. T., Le, K. C., & Nguyen, M. H. (2026). A provincial environmental technology transition indicator for Vietnam: Green finance, spatial diffusion, and absorptive capacity. Environmental and Sustainability Indicators, 32, Article 101554. https://doi.org/10.1016/j.indic.2026.101554

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101554

Keywords: green finance, environmental technology transition, Vietnam, spatial econometrics, institutional quality, absorptive capacity, renewable energy, green patents, spatial spillovers, sustainability indicators, green credit, net zero