No proof that green bonds support emissions reductions, says Bank for International Settlements report

There is no link between green bonds and companies reducing their carbon emissions, according to a report from the Bank for International Settlements (BIS), the Switzerland-based central banking organisation.

Issuer-level ratings based on carbon intensity could be a more effective way to reduce corporate carbon emissions, the report suggests.

In the latest addition to a longstanding debate on whether current green labels and standards contribute to decarbonisation in the real economy, the report assesses whether green bond issuance – which surpassed $250bn in 2019 – is associated with material reductions in carbon emissions at issuers.

The report’s authors used S&P Trucost data to find that “green bond labels are not associated with falling or even comparatively low carbon emissions at the firm level”.

BIS, which published the report but says the views expressed are those of the authors,  launched a green bond fund for its central bank customers last year, which hit the $1bn mark in July 2020. It also warned earlier this year of the risk of “green swan” events.

Firms found by the report to have the highest carbon intensity comprised “virtually equal shares” of green bond issuers and others.

The report’s authors also assessed whether firms reduce their carbon intensity after issuing green bonds, finding “no strong evidence that green bond issuance is associated with any reduction in carbon intensities over time at the firm level”.

There was divergence between sectors however. Industrials and real estate green bond issuers achieved lower carbon emissions than firms that did not issue green bonds, but in the utilities sector, green bond issuers achieved smaller carbon intensity reductions on average.

It noted that, while there are new types of bonds that focus on climate-related outcomes, such as sustainability-linked bonds, climate-aligned or transition bonds, they are still in their infancy. 

Firm-level green ratings could provide a more useful signal to investors and encourage companies to increase their carbon efficiency, the report says, identifying carbon intensity (the ratio of carbon emissions to revenue) as its preferred measure of “greenness”.

The report proposes three “desirable properties of a complementary green rating system” specifically aimed at carbon emissions reduction for firms: 

1. That it provides additional incentives for companies to align with climate goals

2. That it should help investors in decision-making, and

3. That it allows investors and other stakeholders to verify firms’ progress. 

On aligning with climate goals, it argues that “a firm-level rating is better suited to deliver this property than a project-based classification”. 

It recommends using “rating buckets” similar to credit ratings, to make firms easier to compare and motivate them to outperform their market peers. “Granularity in these respects is superior to merely identifying firms as either green or non-green,” it says.

Firms with higher carbon intensity should have more granular ratings, it continues, “because the distribution of carbon intensity is highly skewed. The firms with the highest carbon intensity are also the highest carbon emitters in absolute terms.”

To ensure green ratings are decision-useful for investors, the report emphasises the importance of simplicity, sufficient granularity and stability of ratings. The ‘buckets’ feature, it says, would act as a “coordination device for investors – investor preferences or mandates could be straightforwardly related to rating buckets”. 

It suggests using factual, backward-looking measures like carbon intensity, rather than forward-looking forecasts or more complicated scoring methods. 

The report suggests that “the third parties that construct, collect, and verify carbon emissions data would perhaps be in prime position to supply such a rating at potentially very low cost” – noting that carbon-intensive firms themselves may be unwilling to pay for ratings.