Climate Finance and Green Bond Evolution
Informing Policy with Machine Learning Text Analytics
Download eBook for Free
|PDF file||16.1 MB||
Use Adobe Acrobat Reader version 10 or higher for the best experience.
This research considers the case of financing the response to climate change, also known as climate finance, with emphasis on the labelled green bond market; climate finance is an exemplar of policy challenges in which private sector engagement is integral. This research aims to understand the evolution of themes associated with climate finance and green bonds to identify opportunities to enhance public-private cooperation and facilitate policymaking. The research exists at the intersection of policy analysis, climate science, environmental finance, and machine learning, and makes novel contributions across the data, method, and policy areas. The research employs topic modeling approaches in conjunction with sentiment and qualitative analyses on unstructured data to represent discourse surrounding (1) climate finance, and (2) green and climate bonds. The topic models aid in discovering interpretable, low-dimensional subspaces from corporas extracted from LexisNexis using a crowd-sourced search strategy. In the case of understanding the evolution of climate finance, dominant topics in climate finance news headlines are analyzed temporally and geographically. This is done using an unsupervised probabilistic generative topic model, Latent Dirichlet Allocation (LDA), along with an automated process for model selection and hyperparameter optimization. The LDA climate finance results indicate that topics representing the mobilization of capital and collective action are becoming more prevalent and are regarded more positively in recent years — suggesting a strong case for enhanced public-private partnerships. Labelled green bond opportunities are identified through news and blog articles that correspond to green bond sectors. Sector-specific topics are identified with Correlation Explanation (CorEx), an information-theoretic approach to topic modeling. In the semi-supervised version of CorEx, domain knowledge about the sectors is incorporated via topic anchors. The green bond topic results demonstrate the prevalence of certain investment areas, increasing interest that remains historically high, and market opportunities that may exist by focusing on industry and building sectors and consolidating water and pollution-control sectors. Overall, investments in market structuring and frameworks emphasizing monitoring, verification, and reporting will strengthen transparency and consistency, which will leverage the momentum in climate finance and assists in scaling up the green bond market. Furthermore, the methods and approaches herein have broad applicability to other complex policy settings.
Table of Contents
Motivation And Policy Context: Climate Change, Climate Finance, And Green Bonds
An Economic Framework: Socially Responsible Engagement For Climate Financing
Corpora Creation: A Crowd‐Sourced Search Strategy With NLP
Methods: Unsupervised And Semi‐Supervised Topic Modeling, Hyperparameter Optimization, Evaluation, And Sentiment
The Evolution Of Climate Finance Using Topic Modeling
Green‐Climate Bond Opportunities Using Anchored Topic Modeling
Model Extensions And Future Work: Word Embeddings, Dynamic Topic Models, And Stance With Sentiment
Concluding Discussion: Human Behavior, Ethics, And Social Change
World Economic Forum Global Risks
Greenhouse Gas Emissions By Country And Industry
Conference Of The Parties
Parties And Observers From The UNFCCC
Growth In Green Bonds
World Bank Project Life Cycle
Details Of Belsey And Ghatak's Economic Model
Twitter Related Terms
Common Topic Modeling Implementations
Topic Model Selection Using VEM Versus Gibbs Sampling
Rank Ordered Regions For Climate Finance Corpus
Additional Topic Model Evaluations
Topic Model Interactive Visualization
Sentiment Metric Variations
Explanation of Bonds ‐ A Fixed Income Instrument
Climate Bonds Taxonomy
Types Of Green Bonds
Emotion Category For Green Bond Corpus By Year
Research conducted by
This document was submitted as a dissertation in October 2019 in partial fulfillment of the requirements of the doctoral degree in public policy analysis at the Pardee RAND Graduate School. The faculty committee that supervised and approved the dissertation consisted of Elvira N. Loredo (Chair), Walter L. Perry, and Thomas J. Sullivan.
This publication is part of the RAND Corporation Dissertation series. Pardee RAND dissertations are produced by graduate fellows of the Pardee RAND Graduate School, the world's leading producer of Ph.D.'s in policy analysis. The dissertations are supervised, reviewed, and approved by a Pardee RAND faculty committee overseeing each dissertation.
This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.
The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.