Sovereigns and ESG: Is there value in virtue?

A large residual from the 2010 regression model implies that, based on the country’s ESG score, the sovereign CDS spread should be tighter. To verify this, the sovereign ESG score back in 2010 was recreated, and a regression of this score was examined against CDS spreads. This residual and the spread performance for the period 2010–18 is shown in Figure 3.2. The findings confirm the assumption that large positive residual countries outperform, and spreads tighten more aggressively than for countries that appear rich to the regression. Egypt, Tunisia, Brazil and South Africa were among those flagged as tight and have since widened, while Angola, Latvia and Pakistan were flagged as cheap and are now a lot tighter than in 2010. This suggests that, going forward, one could expect spread tightening from those countries with a large residual as of end-2018.

Note also that out of the 18 countries with residuals over +0.25 in 2010, 17 saw tightening and only one experienced widening

Due to the overlap between the ESG scores and credit ratings, the relationship between the two was also analysed. Unsurprisingly, there was a strong link found between the sovereign ESG score and the average of the three credit ratings for each country. There is also a high correlation between CDS spreads and credit ratings. What is surprising, however, is that the ESG scores were able to explain a large portion of the CDS spread that is not explained by credit ratings.

To examine this, the residual from the CDS spread and sovereign ratings regression were looked at, and these residuals were regressed against the sovereign ESG score. A large proportion of the residual can indeed be explained by the sovereign ESG score – the R² in Figure 3.3 is 73% – therefore, the ESG factors can play an important role in analysing value for sovereign CDS spreads, on top of that already explained by credit rating agencies.

Given the clear link between the sovereign ESG scores and CDS spreads, the next area to investigate was the drivers of the change in ESG and how they compare between EM and DM markets. The drivers of spreads in EM versus DM markets were also examined, as well how much of a role ESG plays in this.

In aggregate, ESG scores increased from 2000 to 2018, with increases seen for all the five different factors. In the DM space, Portugal and South Korea experienced the largest improvement in their overall score, while for EMs it was China and Latvia. Figures 3.4 and 3.5 show what is driving the change in ESG scores for DM sovereigns and EM sovereigns, with both the average DM and EM ESG scores increasing over the period in question. However, and interestingly for DM countries, on average the environmental and social scores improved, while governance scores deteriorated. By contrast, for EM states the environmental, social and governance scores all improved.

Perhaps surprisingly, it was found that DM countries are more impacted by ESG factors than DM countries (see Figure 3.6). This is shown by two different measures. First, the slope of the regression line is much steeper for DM than it is for EM. DM countries’ Also, CDS prices are more impacted by ESG issues than EM countries’ CDS prices – a change in a DM country’s ESG score has a larger impact on its modelled CDS spread than a change in an EM country’s ESG score. In addition, the R² is higher for the DM line of best fit, meaning the variation in ESG can explain more of the variation in DM CDS spreads than EM.

This seems counterintuitive, perhaps because, for EM states, factors other than ESG have a larger impact on their CDS. It may also be that factors such as indebtedness and economic performance have a bigger impact on where EM countries trade than for DM states.

Interestingly, if peripheral or southern European states such as Portugal, Greece or Italy were reclassified as EM states rather than DM, this would serve to flatten the DM curve, putting it more in alignment with the current EM curve. The southern European states have CDS that trade quite wide of where one would expect given their ESG scores – probably because they have “other factors”, including high leverage, that drive their CDS performance.

While the model considers environmental damage, it does not take into account transition risk – the risk that the world takes climate change seriously – as a result of which countries that export fossil fuels (and companies that produce fossil fuel) face diminishing demand for their products. While transition risk is excluded from the model, it is possible to measure a country’s vulnerability by looking at its dependence on oil and how this has changed in recent times. For example, Algeria has seen a reduction in its oil dependence, while both Oman and Bahrain have a high dependency that has been increasing over the last 15 years.

In summary, we were able to produce a simple yet powerful sovereign ESG model, that we applied to 77 sovereigns. The model shows a strong correlation between the sovereigns’ ESG scores and the sovereigns’ 5 year CDS. Moreover, the model appears to have predictive powers: those sovereigns whose CDS appeared cheap in 2010 appeared to outperform those that appears rich in 2010, in the period 2010 to 2018.