Shades of green: Unveiling the impact of municipal green bonds on the environment

Abstract

Green bonds allocate proceeds towards environmentally beneficial projects and sustainable development goals, distinguishing themselves from traditional bonds primarily in the use of proceeds determination. However, investors often find it challenging to assess the carbon reduction potential of these bonds because of the lack of standardised environmental impact reporting. In response to this, our research constructs a unique set of indicators derived from financial and environmental datasets, using multivariate analysis techniques that can accommodate the detection of both linear and non-linear relationships. A novel method combining kernel Principal Component Analysis (kPCA) and Canonical Correlation Analysis (CCA) is applied to detect spatial–temporal cross-correlation in multivariate datasets. This approach handles variable comparability issues and the differential treatment of categorical and numerical variables. A significant finding of this study emerges when this methodology is applied to financial attributes obtained from green bonds issued by municipal agencies (muni bonds), pollution data and environmental (climate) data from nine California counties.The results of the detailed analysis indicate that there is measurable evidence to indicate relationships between green bond issuance and their use of proceeds for pollution reduction efforts. In particular, the results show a clear and interpretable correlation directly linked to the amount of green bond issuance and the effect this is having on pollution reduction, underscoring the tangible impact these financial instruments have on pollution reduction efforts in California.Conversely, when it comes to detecting spatial–temporal relationships between the use of proceeds from green bond issuance and positive climate change effects, this is inconclusive from the current studies’ analysis. It was found that there were weaker cross-correlation relationships observed between climate and green bond financial data set attributes which is perhaps indicative of the fact that climate change effects take a much longer time frame to occur. As such the findings of the analysis in this regard may not indicate that positive climate change effects are not occurring from green bond initiatives, but rather that the ability to measure detectable improvements to climate with regard to the issuance of green bonds is currently limited and will take longer for such effects to manifest in a statistically detectable fashion from the given data. This is particularly likely to be the case given green bonds are only in their infancy as a financial market, having had the earliest issuance only occurring in the last 15-20 years and only substantial growth in the market over the last 10 years. Therefore, this aspect of the research investigating longer-term climate effects with regard to green bond issuance will take longer to develop.