13-07-2022 · Interview

Indices insights: Combing through the climate data forest

We unpack the merits of different carbon metrics used to gauge climate risk in portfolios, including a new climate beta measure. These are discussed with climate strategist, Lucian Peppelenbos, climate data scientist, Thijs Markwat, and Head of Sustainable Index Solutions, Joop Huij, in our roundtable interview.

    Auteurs

  • Lusanele Magwa - Quant Investment Specialist/CPM

    Lusanele Magwa

    Quant Investment Specialist/CPM

Many industry participants – both asset owners and asset managers – are committing to net-zero targets to combat climate change and contribute to the low-carbon transition.1 This transition also triggers the question on how to assess the exposure companies have to so-called transition or climate risk. For instance, to what extent can we rely on traditional carbon emissions data and do we need additional measures to get a clearer picture? We tackle these issues by combing through the climate data forest with our in-house experts Lucian, Thijs and Joop.

Thijs, what are the pros and cons of traditional Scope 1 and 2 carbon emissions data?

Thijs Markwat: “In general, the quality of Scope 1 and 2 carbon emissions data is relatively good. This stems from the clear guidelines outlined in the Greenhouse Gas (GHG) Protocol on how to calculate and report these emissions. The widespread use of this data also ensures that its quality improves over time. Currently there are several data providers that produce and supply carbon emissions data. In our view, such competition is beneficial for the accessibility and quality of the data.”

“But carbon emissions data does have some less attractive features. For instance, the data is backward-looking with possible lags of over two calendar years. As a result, it does not reflect the recent actions taken by companies to reduce emissions. Another disadvantage is that not all companies report their carbon emissions. In general, we see that smaller firms and those from emerging markets have a lower reporting rate. To address this issue, data providers use their in-house models to calculate estimates for companies that fail to report their data. We observe that these estimates can be significantly different between data providers. Thus, this raises the question on their reliability.”

And what about Scope 3 data?

T.M.: “In my view, Scope 3 currently poses one of the biggest carbon accounting challenges. Aside from the issues that plague Scope 1 and 2, there are additional concerns that are specifically associated with Scope 3. Although the GHG Protocol also outlines how to calculate Scope 3 emissions, this is a significantly more complicated task than for Scope 1 and 2. One complexity is that Scope 3 has 15 different subcategories which all require estimates from reporting companies. Moreover, the quality of reported data is often mediocre or even poor. Consequently, the estimates – which are often based on the reported data – are of lesser quality. For example, Scope 3 data can contain large outliers that lead to a single company dominating a portfolio’s carbon footprint.“

“Apart from the data quality, there is also the fundamental problem of double counting. As Scope 3 measures all emissions along companies’ supply chains, longer supply chains typically lead to higher emissions. Therefore, Scope 3 emissions are less transparent, and as a result, less comparable between companies. There are currently no good methodologies available to resolve this double counting issue.”

Joop, what led you down the road towards your climate beta research project?

Joop Huij: “While climate risk has become a key issue for most investors, there are inherent difficulties in addressing it, some of which are data-related. Carbon emissions data gives a good indication, but it is not designed to gauge transition risk. Climate value at risk (VaR) measures – as developed by various data providers – are based on complex, opaque models and produce divergent results between different providers depending on their assumptions. Against this backdrop, we embarked on a project which was based on proposing a solution to these issues. Specifically, we introduced a measure of transition risk exposure that is determined by how an asset’s return is correlated with a climate risk factor, namely climate beta.”

“Climate beta is not based on a model with many assumptions. Its premise is that the market is relatively efficient in pricing in near and medium-term transition risk. Climate beta incorporates current market-based information (no two-year lag) and market expectations, thus making it a more actual and forward-looking measure. It has wide coverage, as market data is available for virtually all listed companies, and has high transparency, as it is not dependent on voluntary disclosures. And perhaps its most attractive feature is that it not only identifies companies that are vulnerable to the low-carbon transition (climate laggards), but also identifies those that are expected to benefit from it (climate leaders). This is something generic emissions data is not able to do.”

In summary, what does your climate beta research paper entail?

J.H.: “In our paper,2 we articulate the climate beta3 concept – which was initially termed ‘carbon beta’ – in detail. Climate transition risk metrics can be difficult to backtest given the lack of historical data and established methods to define when climate risk is high. In our research, we look to overcome this issue by creating a series of tests which can capture periods of climate policy uncertainty and climate change opportunities. We then outline how an asset with a negative climate beta tends to have high returns when climate policy uncertainty increases, and vice versa.”

“In our analyses, we observe robust associations between climate beta and emission footprints and intensities, and green innovation scores. We also see that stocks with higher climate betas have lower returns in periods of elevated climate risk, and vice versa. Furthermore, we look into green innovation and find that within industries, companies with lower climate betas are usually more active in patenting green technologies. This relationship, however, does not exist when one analyzes a firm’s emissions or emission intensity. Therefore, we believe climate beta is an additional climate risk indicator above and beyond traditional carbon data.”

Subscribe - Indices Insights

Receive an update as soon as a new article is available with insights about sustainability, factors or markets.

Aanmelden Read more about Robeco Indices

Lucian, what are main benefits of adding carbon beta to your climate analytics toolkit?

Lucian Peppelenbos: “There is no single metric that can fully capture climate risk. Therefore, we continuously evaluate our climate analytics toolkit and strive to find ways to enhance or expand it. At Robeco, we develop our own fundamental views on the decarbonization pathways for companies in the heaviest emitting sectors, and we complement this with information from external data providers to obtain wider coverage of the investment universe. Climate change is unprecedented, thus we cannot rely on historical data. Key metrics change as climate scenarios evolve. That is why we believe continuous improvements to our toolkit are crucial to help our clients meet specific climate objectives.”

“So when Joop and I discussed the climate beta measure for the first time about two years ago, I was immediately intrigued by the concept. This is because it has a totally different starting point than carbon emissions data and climate VaR metrics. It can, therefore, complement our toolkit. To highlight Joop’s earlier points, what distinguishes climate beta is that it has forward-looking properties, it has broad coverage based on market data instead of self-reported data, and perhaps most importantly, it helps to differentiate climate leaders from laggards.”

Can you define climate leaders and laggards and explain the data challenges in identifying them?

L.P.: “Climate leaders are companies that are expected to benefit from the low-carbon transition. For instance, this could be fueled by green innovation or steps taken towards significant carbon footprint reduction. On the other hand, climate laggards are firms that are expected to be vulnerable to the low-carbon transition. As an example, this could be due to slow decarbonization efforts or business models that are economically linked to a high-carbon economy. All of this is forward-looking, which generic carbon emissions data is not designed to measure.”

“Green innovation provides a good reference point in identifying potential climate leaders. The climate beta research paper outlines how green innovation is driven primarily by the energy, materials and utilities sectors. Within these industries, the more innovative firms have lower climate betas, although they do not necessarily have lower carbon emissions or emission intensities. A good example is a US multinational that is active in the field of power generation. Despite ranking in the top 5% of GHG emitters in our sample, it is also one of the top three issuers of green patents. Thus, its climate beta is among the lowest in the industrials sector. On the other hand, some companies that look ‘clean’ based on their low reported emissions data actually have a high climate beta. For instance, some smaller firms that provide services to the oil & gas sector do not have high-carbon emitting operations, but they are economically vulnerable to a low-carbon transition.”

Thijs, what is the biggest value-add of the climate beta measure in terms of climate risk insights?

T.M.: “This is an easy one for me. One of the biggest problems that affects carbon footprint metrics is that funds which invest in companies that offer climate solutions often score poorly on Scope 1, 2 and 3 emissions, particularly Scope 3. Joop recently showed me how the same funds score well on climate beta. In my view, this demonstrates how climate beta better reflects the climate change awareness of these funds. Thus, it is indeed better suited to distinguish between climate leaders and laggards.”

Are you aware of other measures of transition risk, given that it is a hot topic, and how can climate beta play a role?

T.M.: “Yes, there are many different alternative measures of transition risk and this number is continuously growing. But in my view, the main drawback for most of these metrics is the lack of transparency. In most cases, complex models are used to estimate transition risk. Climate beta, on the other hand, is an intuitive measure and its calculation is straightforward. It simply measures an asset’s sensitivity to transition risk based on market return data.”

“What I also like about climate beta is that it does not use carbon emissions as a direct input, like climate VaR or implied temperature rise metrics. If emissions are a direct input, we typically see that high carbon-emitting companies score poorly on such transition risk metrics. However, some high carbon-emitting companies can be well positioned for the transition given that they contribute positively to the shift towards a greener economy. Climate beta, on the other hand, is not directly dependent on emissions and is therefore better suited to identify these companies.”

Joop, how will you implement this climate beta measure in Robeco sustainable index solutions?

J.H.: “We will impose an index-level restriction on our SDG Low-Carbon Indices – launched in December 2021 – which will result in meaningfully lower climate betas (transition risk) for our sustainable indices versus their respective market cap-weighted indices. The impact of this restriction is particularly noticeable in the weights of index constituents with very low climate betas (more weight) and high climate betas (less weight). We believe this makes our sustainable indices less exposed to transition risk which should be helpful in times when investors are especially concerned about climate policy uncertainty. We can also develop customized indices for clients that place more emphasis on climate risk (climate beta and carbon footprint reduction) to better position their portfolios for the transition.”

Lucian, are you looking to rollout climate beta more broadly in Robeco and do you have examples?

L.P.: “We are currently exploring this. As an example, climate beta can potentially enhance Robeco’s proprietary SDG framework by providing additional information that can be used for determining SDG scores for companies, particularly for climate-related SDGs. It can be used as a screen or an idea generator for fundamental research. We are also investigating if we can use the climate beta measure as a climate risk input in our investment processes alongside other climate metrics.”

Joop, how else can clients implement or benefit from the climate beta measure?

J.H.: “We will estimate the measure based on a specifically designed climate risk factor index. This index tracks the return difference between a basket of ‘polluting’ stocks and a basket of ‘clean’ stocks. In addition to using climate beta to lower the climate risk for our sustainable indices on an index level, investors could also design portfolios or swaps based on the climate risk factor which could act as an overlay to lower the transition risk of their overall portfolios. Going forward, we expect to experience times of heightened climate policy uncertainty and believe that climate beta could be used as an input to temporarily hedge portfolios during those periods. We also believe that the measure could be interesting for investors with business models that are exposed to transition risk and who would like to assess and mitigate this exposure.”

Table 1 | Breakdown of climate metrics

Table 1 | Breakdown of climate metrics

Source: Robeco

Footnotes

1 Robeco, March 2022, “2022 global climate survey”, Robeco publication.
2 Huij, J., Laurs, D., Stork, P. A., and Zwinkels, R. C. J., November 2021, “Carbon beta: a market-based measure of climate risk“, SSRN working paper.
3 The methodology and calculation of the climate risk factor and climate beta are explained in detail in the above research paper.