Voluntary carbon credits index @Royalton Partners

My role

At Royalton Partners (Luxembourg), I developed a POC for a systematic index tailored to the voluntary carbon credits market. This initiative aims to establish the market’s first benchmark index, which is not available as open-source.

What are voluntary carbon credits?

Voluntary carbon credits are a type of tradable certificate that represents the reduction, removal, or avoidance of one metric ton of carbon dioxide (CO2) or its equivalent in other greenhouse gases from the atmosphere. Unlike compliance carbon credits, which are often mandated by government regulations and emission reduction targets, voluntary carbon credits are purchased and retired voluntarily by organizations, businesses, or individuals to offset their carbon footprint.

Why?

With the anticipated growth of the voluntary carbon credits market, there is a pressing need for a benchmark index. Such an index is crucial for providing a comprehensive view of market trends and health, enabling companies to align their carbon offset strategies with broader environmental goals. Additionally, fund managers will leverage this index as a reference point for performance assessment.

How?

Given the diverse nature of carbon credits and the decentralized state of market data, traditional indexing methods like the Laspeyres index fall short in accurately capturing the price dynamics of carbon credits. Price variations are significantly influenced by the characteristics of the carbon credits themselves. This challenge is akin to pricing in the housing market and can be addressed through hedonic regression techniques, which consider the qualitative attributes of assets in their valuation, see for ex:

• Lin, Min-Bin and Wang, Bingling and Bocart, Fabian and Hafner, Christian M. and Härdle, Wolfgang Karl, 代 DAI Digital Art Index: A robust price index for heterogeneous digital assets (2022). Available at SSRN: http://dx.doi.org/10.2139/ssrn.4279412
• https://www.insee.fr/en/metadonnees/source/serie/s1028

Challenges

The primary challenge lies in constructing a reliable data pipeline that remains unaffected by external changes, such as modifications in data formats or the introduction of new data sources. The process involves:

• Aggregating data from various sources.
• Parsing, correcting, cleaning, and standardizing data into a unified database with predefined characteristics of various types (numerical, ordinal, categorical, etc.)
• Implementing feature engineering practices.
• handling missing values (for example via clustering and classification for categorical variables), new categories.
• Conducting thorough data quality assessments and generating reports.

Additionally, the volatile nature of the carbon credits market and its propensity for outliers pose significant challenges, necessitating a refined approach to hedonic regression that minimizes outlier impact and ensures a stable index.

My contributions

My role encompassed several activities:

• Scraping and parsing unstructured data.
• Cleaning data to ensure consistency and accuracy.
• Developing and maintaining a robust ETL data pipeline.
• Formulating a proprietary hedonic model, fine-tuning its parameters, and conducting error analysis to correct for outliers.
• Validating the model through rigorous backtesting.