Five years ago, on 12 December 2015, nearly 200 countries committed to achieving the Paris Accord’s climate goals, including one of its three main objectives: “making finance flows consistent with a pathway towards low greenhouse gas emissions and climate-resilient development” (United Nations 2015). Green finance instruments have a key role to play in mobilising private capital, and green bonds are perhaps the most successful of these instruments thus far. Global issuance has been growing rapidly in recent years and in 2019 surpassed $250 billion, or about 3.5% total global bond issuance ($7.15 trillion).
On the investors’ side, an increasing number of private and public sector entities have explicit mandates for portfolio allocations that support the mitigation of climate change. These considerations have grown – including in the central banking community, through the Network for Greening the Financial System (NGFS) – as the link between climate change, systemic risk, and financial stability has become clear (Fender et al. 2020, Bolton et al. 2020).
A context in which investors’ preferences are evolving and demand for ‘green’ instruments is rising makes it even more important to clarify taxonomies and ensure the ‘greenness’ of proposed investment alternatives. In a recently published article (Ehlers et al. 2020), we assess whether corporate green bonds1 are issued by less carbon-intensive firms, and if green bond issuance leads to a reduction in carbon intensity of the issuer. We show that green bonds have so far been disappointing in both respects. We then outline desirable properties for a rating system that could help spur the mitigation of climate change.
What green bonds are designed to deliver
Green bonds differ from ‘normal’ bonds in only one (non-financial) aspect: they carry a green label.
The green label is a signal that the proceeds of the bond are used for environmentally beneficial projects. The range of environmental benefits of green bonds is not limited to climate change mitigation (e.g. reduction of greenhouse gas emissions) and covers a broader range of benefits such as climate change adaptation, water security, or waste reduction.
While there are no global regulatory minimum standards for green bonds – or how environmentally beneficial projects are defined (Ehlers and Packer 2017) – the private sector has developed the high-level green bond principles (GBPs) as well as more specific voluntary ‘standards’ based on the GBPs (e.g. CBI 2019). A major development in setting official standards for green bonds is the proposed EU green bond standard (EU TEG 2019a), based on the EU sustainable finance taxonomy (EU TEG 2019b).
However, even if issuers were to strictly adhere to the currently voluntary standards, these standards do not ensure beneficial environmental outcomes. Hence, green bonds should not necessarily be expected to lead to a reduction in carbon emissions at the firm level. We argue that this is desirable if green bonds are to help achieve the financing of green projects and, at the same time, of high-level climate goals.
Green bonds and carbon intensity of firms
Our preferred measure of firm ‘greenness’ is carbon intensity, or the ratio of carbon emissions to revenue. Unlike a simple absolute measure of carbon emissions, this ratio measures the firm’s carbon efficiency. Firms that use greener technologies and energy can achieve lower carbon emissions at the same level of economic activity and size. Our sample is based on around 16,000 listed firms (>99% of global market cap) using S&P TruCost data.
Less informed investors might expect firms with very high carbon intensities to be disqualified as issuers of green bonds. However, Graph 1 indicates that for Scope 1 (direct emissions emanating from the firms’ own resources) and Scopes 1-2 (Scope 1 plus indirect emissions from consumed energy), a greater fraction of green bond issuers than other issuers have a carbon intensity above 100 tonnes of CO2 per million dollars of revenue. For Scopes 1-3 carbon intensities (Scopes 1-2 plus all indirect emissions from upstream and downstream activities), firms with the highest carbon intensity comprise virtually equal shares of green bond issuers and others.
Figure 1 Distribution of carbon intensities: Green bond issuers versus other firms, FY2018
Comparing the carbon intensities of green bond issuers with those of other firms buttresses two important points previously made in Ehlers and Packer (2017). First, even if bond proceeds flow into green projects (e.g. renewable energy), issuers may be (and often are) heavily engaged in carbon-intensive activities elsewhere (e.g. coal power plants). Second, the wide range of varying green bond standards allows a very broad assortment of firms to issue green bonds, each deemed to be green for different reasons.
There is also no clear evidence that green bond issuance is associated either with a reduction in carbon intensities over time at the firm level (Graph 2), or with lower intensities than firms that did not issue green bonds (see Ehlers et al. 2020).2 Around 60% of green bond issuers in our sample show a reduction in Scope 1 carbon intensities after three years; but only about 30% show a reduction when looking at broader Scopes 1- 3 intensities.
Figure 2 Green bond issuance and changes in issuers’ carbon intensity
Desirable properties of a complementary green rating system
What are the desirable properties of a rating system to support the transition to a low-carbon economy? We highlight a few key properties here and refer to our article for a more comprehensive discussion.
We argue that a firm-level rating is better suited to deliver this property than a project-based classification. The firm is the decision-making unit when it comes to the carbon footprint of economic activity: that is, the production process, choices of inputs, outputs and means of distribution.
Further, rather than focussing on firms in the lowest end of the carbon intensity distribution, green ratings should focus on the activities of high polluters, as their improvement will yield the highest gains in the overall reductions of carbon emissions. In fact, the 1% of firms with the highest carbon intensities (>99% percentile) are responsible for close to 40% of aggregate Scopes 1-3 emissions (Figure 3).
Figure 3 Absolute carbon emissions by carbon intensity percentile, FY2018
The rating of ‘greenness’ should be based on simple verifiable outcome-based measures, such as the carbon intensity of firms. This is an easy measure for potential investors to understand, and its availability across a wide range of firms allows for straightforward verification. To avoid firms exploiting loopholes, including by outsourcing their carbon-intensive activities, we propose using the broadest possible range of scopes (currently Scopes 1-3), which covers indirect emissions from the inputs used as well as distribution and usage of products. By doing so, we can achieve a win-win wherein green projects can be financed while firms are given more incentives to reduce their overall carbon intensity.
Example of a firm-level carbon-intensity rating
As an illustration, we develop a simple example of a firm-level carbon-intensity rating (Table 1). One of its key features is the higher granularity of rating buckets for firms with high carbon intensities, which gives those firms a better chance and stronger incentive to improve their ratings – more so than would a binary green rating or certification. The carbon intensity cut-off points of our rating are fixed over time and allow firms to gradually improve, in the spirit of high-level climate goals.
The simple illustrative rating system has further beneficial properties, such as a certain degree of stability of ratings over time (a key requirement for investors) and simplicity. It would be easy to implement and hence low-cost.
Table 1 Green rating: Carbon intensity cut-off points
Green bonds have played an important role in developing the market for sustainable finance instruments. Its rapid growth is a testament to investors’ demand for such instruments, which is likely to grow further. Complementary ratings to current green certifications could help to convert market trends into much needed reductions in carbon emissions.
A simple rating system based on firms’ carbon intensities, such as the one discussed here, could help to inform investor allocation decisions not only for bonds but also for stocks. Various other related efforts are under way, such as standardised impact reports for green bonds, which could include an assessment of achieved (or expected) carbon reductions. New types of bonds – such as sustainability-linked bonds, climate-aligned or transition bonds – also focus on outcomes, including carbon emission reductions (CBI 2020, ICMA 2020). But such efforts are still in their infancy. In order to contribute to reducing the huge risks related to climate change, we need to improve information and clarity for potential investors.
Bolton, P, M Despres, L Pereira da Silva, F Samama and R Svartzman (2020), “The green swan”, Bank for International Settlements and Bank of France.
Climate Bonds Initiative (2019), “Climate Bonds Standard V3.0”, December.
Climate Bonds Initiative (2020), “Climate Bonds Taxonomy”, January.
Ehlers, T and F Packer (2017), “Green bond finance and certification”, BIS Quarterly Review, September, 89–104.
Ehlers, T , B Mojon and F Packer (2020), “Green bonds and carbon emissions: exploring the case for a rating system at the firm level”, BIS Quarterly Review, September, 31-47.
EU Technical Expert Group on Sustainable Finance (2019a), “Report on EU Green Bond Standard”, June.
EU Technical Expert Group on Sustainable Finance(2019b), “Taxonomy: Final report of the Technical Expert Group on Sustainable Finance”, March.
Fender, I, M McMorrow, V Sahakyan and O Zulaica (2020), “Reserve management and sustainability: the case for green bonds?”, BIS Working Papers, no. 849, March.
United Nations (2015), “Paris Agreement”, Article 2.1, December.
1 Our study is solely focused on corporate bonds. Sovereign green bonds are another important segment of the green bond market.
2 The standard deviations are several multiples of the mean changes presented in Figure 2 and hence not shown.