Carbon Premium around the World

Apart from COVID-19, few topics garner as much global attention these days as the looming climate crisis. And let us not forget that before the outbreak of the pandemic the front pages of most newspapers were filled with stories and pictures of the extraordinary wildfires in Australia, which by some estimates have cost the country up to 5% of GDP. The visible warming of the planet (2019 was the warmest year on record) has put climate change at the centre of people’s preoccupations all over the world. Yet, even a casual observer can see that countries have implemented widely differing policies to combat climate change. Some regions of the world, most notably Scandinavia, have been at the forefront of transitioning their economies towards renewable energy. Other countries have taken very limited steps in this regard, or even reversed past policy commitments towards reducing carbon emissions, as the United States has done by brashly pulling out of the 2015 Paris Agreement. Similar antagonistic stances towards climate change mitigation have been appallingly on display by the Australian government in the midst of the worst wildfires in Australian history and by the Brazilian government facing a resurgence in forest fires in the Amazon by wildcat miners and farmers.

How are the growing awareness about climate change and the different climate policy stances across countries reflected in financial markets? Are investors factoring in a carbon premium in their stock valuations, and if so how does this premium vary across countries? These are the questions we address in this study by looking at how stock returns in 77 countries for more than 14,000 publicly listed companies are affected by the risk associated with carbon emissions. Two leading hypotheses suggest themselves. One possibility is that financial markets are globally integrated, so that the same climate change risk in Bangladesh or in New York is priced the same. The other is that there is substantial segmentation in financial markets, so that the pricing of carbon risk in each country mostly reflects local conditions.

Investors in companies that supply fossil fuel energy, and in companies that rely on this energy for their operations, are increasingly exposed to risk with respect to policies seeking to curb carbon emissions and to technological risk from alternative, more and more affordable, renewable energy. Using firm-level carbon emission and financial data we quantify the carbon premium, that is, the return that compensates investors for taking on the transition risk.

If transition risk is disregarded one would expect to see no significant correlation between stock returns and CO2 emissions (once one controls for all other known risk factors and firm characteristics). Yet, we find that carbon emissions do affect stock returns in most geographic areas of the world. The premium is economically sizable with respect to both direct emissions (scope 1), indirect emissions from consumption of purchased electricity, heat, or steam (scope 2), and other indirect emissions from the production of purchased materials, product use, waste disposal, outsourced activities, etc. (scope 3). Interestingly, what matters is both the total level of CO2 emissions produced by companies as well as their year by year changes. The former measure can be understood as a long-term risk projection, given that emissions are highly persistent, while the latter is a short-term projection of the risk. However, we do not find any premium associated with emission intensity (the ratio of emission level to sales), which is a common measure used by institutional investors. This result suggests that absolute levels of carbon emissions are more important metrics of carbon risk, which makes sense given that carbon emission reduction policies are based on levels and not on intensity.

We further explore which factors carbon premia are associated with across countries, industries, and firms, and discover a number of interesting patterns in the data. The first surprising result is that the level of economic development does not explain cross-country variations in the carbon premium. On the other hand, several other country characteristics matter. Both “voice” (how democratic political institutions are) and “rule of law” significantly affect the carbon premium associated with changes in emissions. More democratic countries with stronger rule of law tend to have lower carbon premia, other things equal. One possible interpretation of this result is that in these countries green public opinion has already resulted in significant tightening of regulations of carbon emissions, so that the transition risk going forward is lower. In support of this interpretation, we also find that the carbon premium is lower in countries with a higher share of renewable energy, and higher in countries with larger oil, gas, and coal extracting sectors. We further find that countries that have been exposed to greater damages from climate disasters (floods, wild-fires, droughts, etc.) have a somewhat higher carbon premium associated with the level of direct emissions. This is not really surprising in light of the fact that climate disasters tend to raise awareness about climate change.

Given that climate change has become a salient issue for investors only recently we also explore how the carbon premium around the world has changed in recent years. We do this by comparing the estimated premia for the two years leading up to the Paris agreement and following the agreement. A number of striking results emerge from this analysis. First, when we pool all countries together, we find that there was no significant premium before the Paris agreement, but a highly significant and large premium in the years following the agreement. This general result is consistent with the view that investors have only recently become aware of the urgency of climate change.

Given that many new firms are added to our sample in 2016, one possible explanation we explore is that the rise in carbon premium is entirely driven by the addition of new firms. We find that this is not the case, after estimating the premium again post Paris on the smaller sample that excludes the new firms. Second, when we break down the change in the carbon premium around the Paris agreement by continent, we find that the premium is insignificant in North America before and after Paris, has declined in Europe, but, astonishingly, has sharply risen in Asia. In effect, Asia is entirely responsible for the rise in the global carbon premium around the Paris agreement.

One obvious mechanism that can give rise to a carbon premium is divestment. We systematically explore this channel by looking at the extent to which institutional investors are underweight companies with high carbon emissions around the world. We make several remarkable discoveries. First, while there is significant divestment, it is all based on a direct emission-intensity screen. Neither the level of emissions (whether direct or indirect) nor changes in emissions significantly affect institutional investors’ holdings, with the exception of hedge funds, who can generally be seen as holding contrarian positions from the other institutional investors. This is true around the world, with significant divestment in the U.S. and in Europe of high carbon intensity firms. There is also significant divestment in Asia, but to a lesser extent than in Europe or the U.S, and relatively little divestment in China.

Second, we break down divestment policies by different institutional investor categories. We find that overall divestment is concentrated among three categories, investment companies, independent advisers, and pension funds, who are significantly underweight companies with high direct emission intensity. The most revealing breakdown, however, is between foreign and domestic institutional investors. Domestic institutional investors concentrate their exclusionary screening primarily on foreign companies. Another interesting breakdown is between passive and actively managed funds. To the extent that passive investments are index investments there would appear to be no scope for divestment. However, in recent years index providers have put in place low-carbon index alternatives to the major market indexes. By holding the low-carbon version, passive investors thus can reduce their exposure to carbon emissions. What we find is that, although most divestment is concentrated among active investors, there is also significant divestment among passive investors.

Overall, our analysis paints a nuanced picture of the pricing of carbon transition risk around the world. The pricing is uneven across countries but widespread in North America, Asia, and Europe. The pricing is also rising, with a significant increase post Paris agreements. There does not appear to be a direct relation between the pricing of transition risk and divestment. Although there is significant divestment related to carbon emissions, it is not directly related to transition risk.

The complete paper is available for download here.