Analysing the link between environmental performance and sovereign credit risk

ABSTRACT

This study investigates the effects of a country’s environmental performance on sovereign credit risk. We use sovereign credit default swap (CDS) spreads for a large panel of countries to analyse whether the relationship between environmental performance and credit risk varies by the maturity of the credit instrument and by the level of fiscal performance. Using a dynamic panel generalized method of moments model, we document a negative association between environmental performance and credit risk. This relationship does not exhibit substantial variation by CDS maturity.

KEYWORDS:

• Environmental performance
• sovereign credit risk
• dynamic panel GMM
• CDS

JEL CLASSIFICATION:

• G15
• F30
• Q51

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ http://www.longfinance.net/images/reports/pdf/unepfi_erisc_2012.pdf.

² https://www.moodys.com/sites/products/ProductAttachments/Climate_trends_infographic_moodys.pdf.

³ Soytas, Denizel, and Usar (2016) provide a detailed review of the literature on sustainability and firms’ financial performance.

⁴ Appendix A provides a list of the countries in the sample, as well as subsamples based on fiscal balance. Our sample of countries is based on the intersection of EPI and CDS data availability for all years in our sample.

⁵ The IMF and World Bank also cite International Financial Statistics, Government Finance Statistics Yearbook, and OECD as data sources.

⁶ The total reserves/GDP value are multiplied by 100 in the analysis.

⁷ The source of the environmental performance index EPI and environmental risk exposure is the Yale Centre for Environmental Law and Policy (YCELP) at Yale University, Yale Data-Driven Environmental Solutions Group at Yale University.

⁸ Some of the performance indicators used to estimate EPI scores changed in 2017. The inability to backcast the EPI estimation using data for 2017 and beyond, caused us to limit the study period to 2016.

⁹ https://epi.envirocenter.yale.edu/.

¹⁰ When the ratio of the variance of the panel-level effect to the variance of the idiosyncratic error or the autoregressive parameters is too large, the Arellano and Bond (1991) approach can perform poorly. To address these issues, Blundell and Bond (1998), expanding the work of Arellano and Bover (1995), developed a system GMM estimator that considers these issues by expanding the instrument list to include instruments for the level equation.

¹¹ Year dummies have been in the system GMM to capture the influence of aggregate time‐series trends. They should capture the effect of any year variation of the dependent variable across the panels. The year dummies are significant for 2009, 2012, 2013, and 2014 in the European subsample. Year dummies were also found to be significant when examining the full sample.

¹² There are insufficient observations to carry out system GMM analysis using only commodity-dependent countries in our sample.