Green Bond Study

The Green Bond market has increased exponentially since the first issuance in 2007. Since then, green bonds have attracted a wide range of buyers, from agnostic to ESG-focused investors. This rise in interest has put focus on the characteristic pricing of green bonds, questioning whether the “green” feature entails a yield premium with respect to conventional bonds. The purpose of the paper is to study the green bond yield premium. This refers to green bonds being priced above or below conventional bonds of similar characteristics. Although active investors should not be dissuaded by a negative premium – as it can be offset by future excess returns, market participants have been wondering to what extent green bonds price differently than the rest of the bond market.

From an issuer’s point of view, a green bond issuance is more expensive than a conventional issuance due to the need for external review, regular reporting, and impact assessments. From an investor perspective, financially speaking, there is no fundamental difference between a green and a conventional bond; green bonds ranking pari passu with similar bonds have no additional rights for the underlying project. Hence, a green premium should be considered as a market anomaly. The paper studies this issue using two different methodologies:

A Top-Down approach: An overall green portfolio is compared to an equivalent conventional bond portfolio.
A Bottom-Up approach: Takes the definition of an intra-curve green bond premium, comparing a green bond to a hypothetical conventional bond of the same issuer, currency, and seniority.

The Top-Down Approach

In the first method of analysis, we compare a “Green Index” (the Bloomberg Barclays MSCI Global Green Bond Index) to a conventional bond index, or “benchmark” (the Bloomberg Barclays Global Aggregate Bond Index) from September 2016 to September 2020. To compare the performance of both portfolios, the “benchmark” is weight-adjusted to mimic the currency, sector, credit quality and maturity features of the Green Index.

The results show that on average, green bonds have a negative premium. The mean value being -4.7 bps, and the maximum value -2.5 bps. The premium was the lowest, -10.5 bps, during the first weeks of the COVID-19 crisis.

Additionally, comparing the performance of both portfolios highlights an identical Sharpe Ratio over the period despite lower excess returns of the green portfolio, which is due to its lower volatility. It is interesting to highlight the fact that the green portfolio outperformed during two periods of extreme stress: 2018 and 2020. This could be explained by:

The nature of green bond issuers: historically large institutions with well-established governance structures, hence more resilient. E.g. Multilateral Development Banks.
The nature of green bond investors: mainly large pension funds and insurance companies with buyand-hold strategies. i.e. less likely to move away from investments in times of crisis.

A breakdown by currency, time-to-maturity, sector, and credit rating shows the following:

Euro-denominated green bonds (majority of portfolio) have a negative premium with an average of -7.3 bps. Meanwhile, USD-denominated green bonds fluctuated from a 10bps premium in 2018 to -21bps during the recent crisis.
Only maturities between 5 and 10 years have a significant negative premium. Most of the time, premia of lower maturity trend above premia of longer maturities.
All sectors except industrials and local authorities exhibit negative premia, although they are only significant for financials and agencies.
All sectors except industrials and local authorities exhibit negative premia, although they are only significant for financials and agencies.

The green premium defined in this method includes a potential “green bond issuer premium” often referred to as “green halo” i.e. the lower yield an issuer gets on all its bonds from having issued some as green bonds.

The Bottom-Up Approach

In this case, each green bond is when possible, compared to two conventional bonds. Thus, the spread of the green bond is compared to the spread of a theoretical interpolated conventional bond of the same duration. The green premium will therefore be the difference in spread between the green bond and a comparable bond with the same issuer, currency, seniority, and modified duration. Compared to the first approach, this method only focuses on the intra-curve green premium and intends to assess the difference coming solely from the green bond format. The weekly data is aggregated over the period April 2019 to September 2020.

The results show that on average, the green bond premium is negative (mean of -2.2bps) and with a high number of outliers (+/- 4 sigmas), exacerbated by the effect on liquidity of the COVID-19 crisis.

A breakdown by currency, time-to-maturity, sector, and credit rating shows the following:

Both EUR and USD green bonds have a statistically significant negative premium (respectively -1.6bps and -3.7bps) USD premium is also twice as volatile.
Within Credit, Utilities show the more negative and statistically significant premium (-3.9bps).
Finally, we find that that the premium and its significance increases with the ESG quality of the issuer, i.e. beyond the use of proceeds, green bond investors reward a more negative premium to issuers with better extra-financial standards at the company level.

We also perform a linear regression of the green bond premium against the following variables: age, duration, size, currency, seniority, sector, country, certifications, and Amundi E-ratings. They find that:

Domestic bonds and bonds certified by the CBI1 tend to show low premia.
All credit rating categories exhibit negative premium, with their mean ranging from -3.13bps for A-rated issuers and -1.38bps for Aa-rated issuers.
All regions exhibit a negative premium but only Europe’s is statistically significant.
All sectors except sovereigns exhibit a more negative premium than the financial sector.
The premium lowers with smaller sized bonds as well as with age although not significantly.

Finally, to check whether the different liquidity features of the green bond market can explain part of the green bond premium, we filter the universe to retain bonds with similar liquidity characteristics. They find that the green bond premium is not a consequence of a liquidity premium, as the premium is even lower after neutralizing liquidity.