Executive Summary
Since its founding in 1994, the GMO Emerging Country Debt team has incorporated ESG-related factors into its systematic sovereign risk assessment process and has engaged in discussions on ESG matters in due diligence meetings with sovereign creditors, multilaterals, and other organizations.1 In this paper we summarize our more recent efforts to bolster that process through the use of proprietary ESG indicators. We begin by reviewing the implicit role that ESG has traditionally played in our process. We then discuss some of the unique challenges involved in including ESG as a factor in sovereign emerging country debt investing. Next, we describe our recently adopted strategy for systematically involving ESG in our analysis in light of those challenges. Finally, we discuss the results of our analysis to date, as well as considerations for future research.
What Role Has ESG Traditionally Played in Emerging Sovereign Debt Investing at GMO?
The aim of our sovereign risk assessment process has been, and continues to be, to compare emerging countries to one another on the basis of their economic fundamentals, toward identifying which appear "rich" or "cheap" in terms of their sovereign spreads at any given point in time. We have traditionally approached this task by distilling various economic variables into three main "pillar" scores – Economic Structure, Fiscal Sustainability, and External Liquidity – each of which is used as a final input toward establishing a single score for credit quality. The main output of that process is shown in Exhibit 1, which will be familiar to many of our investors, plotting countries' fundamental credit scores derived from this analysis on the x-axis, and their observed 10-year sovereign Z-spreads as of December 2020 on the y-axis.
EXHIBIT 1: COMPARING SPREADS VS OUR OWN ESTIMATES OF SOVEREIGN RISK VIA OUR TRADITIONAL APPROACH
Source: GMO
ESG factors have, since the beginning, had a presence in that process via two main channels: 1) through its factor-based component, and specifically through the presence of the World Economic Forum's Global Competitiveness Index (GCI) within our "Economic Structure" pillar; and 2) through qualitative considerations that complement that factor-based component.
The inclusion of GCI as a factor-based input has been aimed at measuring the strength of countries' economic, social, and governmental institutions, with the goal of assessing the structural health of their economies. GCI is a composite of dozens of individual variables, some of which have clear ESG relevance. Some examples of the variables within GCI that could be considered ESG in nature would be the presence of energy-efficiency regulation and environment-related treaties within the "Environmental" category; homicide rate, life expectancy, and mean years of education within "Social"; and judicial independence, press freedom, and incidence of corruption within "Governance."2
Secondarily, ESG considerations often play a significant role in the supplemental qualitative research that accompanies our factor-based approach. That qualitative component involves country-level information that is exogenous to our fundamental model, often directly gleaned from our regular engagement with sovereign issuers, multilateral organizations, or third-party analysts. For instance, even as the historical volatility of growth and inflation are explicitly included in our systematic process, we have historically taken into account countries' vulnerability to natural disasters (that is, environmental risk) as a further consideration when assessing a "fair" sovereign spread. The same goes for other risks that can be considered social or governance-related in nature – for instance, a country's level of social cohesion and the risk of civil strife (social), or the risk of an adverse change in economic policies following an upcoming election (governance). Until recently, these phenomena have not been explicitly included in our systematic process, but nonetheless have played an important role in our assessment of countries' creditworthiness and, ultimately, our investment decisions.
Some ESG-related Challenges as They Relate to Sovereign Risk Analysis
Recently we began exploring potential ways to involve ESG factors more directly in our sovereign risk analysis process beyond their presence via GCI and the qualitative component mentioned above. We recognized numerous sources of complexity in working to that end and highlight three in particular below.
The measurement of countries' ESG quality is not straightforward: We see considerable ambiguity in how ESG variables are measured and accounted for. Our traditional inputs into our systematic process are relatively straightforward; per capita income in U.S. dollars, for instance, is direct in terms of measurement and unambiguous in its relationship with sovereign risk (richer countries tend to be more creditworthy). By contrast, estimating a country's relative strength along the parameters of environmental, social, or governance issues is much less straightforward. One notable example is the incidence of corruption, a commonly cited component of countries' governance ratings, which can be measured or proxied in numerous ways depending on the preferences of the analyst and its intended use.3
Further, we see no clear consensus within the asset management industry on how to handle ESG-related phenomena that are undesirable from a global vantage point, but not necessarily at the country level. One example here is greenhouse gas emissions: fossil fuel-producing countries (including much of the emerging world) privately benefit from the existence of these extractive industries, yet at present these countries do not fully bear the public costs at the global level, even amid incipient coordination efforts such as the Paris Climate Accord. Thus, at this point in history, assigning a negative value for such countries in terms of our own sovereign risk assessment or our ultimate investment decisions may be morally justifiable, but not necessarily optimal from the perspective of gauging countries' creditworthiness and maximizing portfolio returns. Indeed, emerging economies tend to be much more reliant on fossil fuel production than their developed counterparts, making this terrain particularly rugged for analysts of emerging debt (see Exhibit 2).
EXHIBIT 2: FOSSIL FUELS PLAY A LARGER ROLE IN EMERGING ECONOMIES
Source: UNCTAD, J.P. Morgan, IMF
Countries' ESG quality appears to be positively correlated with income, and negatively correlated with yields. Another thing we observe is that ESG quality is highly correlated with income levels, meaning that emerging countries tend to be, all things being equal, worse than developed countries on these metrics. Economists have long accepted the existence of the relationship between countries' level of development and things like environmental quality or the strength of political institutions.4 Indeed, Exhibit 3 shows a clear correlation between countries' total ESG scores (using metrics from MSCI) and their level of per capita income. From our vantage point, any analysis of an emerging asset class or individual portfolio's level of ESG quality must take this fact into consideration.
Making matters more complicated is the fact that within emerging country debt, poorer countries not only tend to have worse ESG scores, but also offer higher yields (see Exhibit 4). Thus, all things being equal, portfolios that are overweight higher-yielding sovereigns face an uphill climb in order to match or exceed the average ESG quality of their benchmark, even as they are likely to deliver positive alpha in the long run. Conversely, to disqualify low-scoring ESG countries from eligibility within an emerging debt portfolio, or to otherwise limit their weights in the portfolio irrespective of other factors, runs the risk of skewing the portfolio toward lower yields, and potentially lower returns.5
EXHIBIT 3: ESG SCORES ARE HIGHLY CORRELATED WITH PER CAPITA INCOME…
Source: World Bank, Bloomberg, MSCI, J.P. Morgan
EXHIBIT 4: …AND NEGATIVELY CORRELATED WITH YIELDS
Source: World Bank, Bloomberg, MSCI, J.P. Morgan
We see a high degree of multicollinearity between ESG factors and other variables in our systematic process. As mentioned earlier, the primary focus of our systematic sovereign risk assessment process is to use regression analysis to establish a "fair value" for each country's sovereign spread and compare this to what we observe in the market (that is, whether or not a country is "rich" or "cheap"). Nonetheless, as sovereign analysts we also often look for additional insights at the variable level as we move toward making our final investment decisions. Even without involving ESG factors, a high degree of correlation among inputs into our model sometimes makes these additional insights elusive, but this challenge of multicollinearity is further magnified when we involve ESG.6
Our Approach to Integrating ESG as a Factor in Our Sovereign Risk Assessment Process
We identified three core principles to guide us as we sought to involve ESG more substantially in our systematic process:
- Relevance. One of our priorities has been to establish our own framework for "rolling up" individual ESG-related data into proprietary aggregated scores, rather than relying on third-party, pre-packaged metrics. This strategy allows us to focus on what matters specifically to the emerging sovereign debt asset class, and to our own investment approach.
- Performance. Integration of ESG metrics should, in our view, contribute toward assessing sovereign risk, rather than simply serving as a signal of our values and priorities. To that end, we chose to use a regression approach when establishing a fourth (ESG) pillar score, as we use for the other pillars.7 This approach allows us to avoid arbitrary and subjective judgments about the relative weights of individual ESG-related variables and lets the statistical process itself indicate which variables are helpful to reaching our final goal.8
- Continuity. Our strong preference has been to keep the basic structure of our existing systematic process intact. As mentioned earlier, that process distills countries' economic fundamentals into three pillars, each with its own 1 through 100 score. To that end, we have elected to establish a fourth, ESG-dedicated pillar upon which our final credit score is based.
Continue reading here.
