Sustainable investing implies that green assets entail low expected returns because (i) green investors relish holding them and (ii) such assets hedge climate risk by encouraging pro-environmental outcomes. In a new paper, we evaluate whether these predictions are supported by the data using a sample of green bonds. In the process, we investigate (i) whether green bonds are associated with significant returns for bondholders and equity investors, both at the time of issuance and after; and (ii) whether the carbon emissions of the issuers of green bonds fall post-issuance.
Green bonds represent a noteworthy asset class of their own. The sales of green bonds reached a monthly record of $32 billion in September 2020, bringing the market’s overall size to almost $1 trillion (Wall Street Journal, 2020). Governments and companies issued $600 billion in green debt in 2021, which is almost half the total raised since the inception of this asset class (Financial Times, 2021). However, while investor demand for green bonds has increased, skeptics still question how much they help the environment (The Economist, 2020).
There are different theories in sustainable investing. One is that socially conscious investors are willing to trade wealth for societal benefits, which translates to a positive premium, or lower yields, for green securities. However, another theory is that green bonds should carry a negative or a zero premium, or higher yields because (i) green investing does not generate positive net present value (NPV) per se and (ii) the issuance of a green bond is costly compared with conventional bonds because issuers typically need to submit to a third-party validation that the proceeds will be spent on environmentally friendly projects.
To examine these competing hypotheses, we document the stock market reaction to the issuance of green bonds and the cross-sectional factors (bond, firm, and country characteristics) associated with such reaction. Post issuance, periodic performance reports on green projects and the associated assurance, if any, of green auditors would reveal additional information and would thus be priced in the secondary market. As such, we also examine the risk and the debt market returns of green bonds in the secondary market. Here, we separate the polluting sectors from other sectors to assess any differences in the sectoral premium of such bonds.
The other crucial question is whether green bonds actually lead to environmentally friendly outcomes. The “signaling” hypothesis suggests a costly environmental commitment made by the issuer to long-term sustainability. Using a general equilibrium model, Glamsrod and Wei (2018) argued that green investing can reduce global coal consumption, increase the market share of non-fossil electricity, and reduce global CO2 emissions. While the evidence related to the actual outcomes of green investing is somewhat ambiguous, Flammer (2021) supports the view that the carbon emissions of green bond issuers decrease in the future. However, critics argue that green bond issuers do not always fulfill their pledge to invest the proceeds in “green” projects only because the penalty for reneging on such a pledge is low (Wall Street Journal, 2021). Green bonds have become more popular in recent decades, with the majority issued by firms head-quartered in the U.S., China, Sweden, France, and Malaysia. The corporate issuers of green bonds mainly belong to the financial, energy, utility, and industrial sectors.
Using a sample of 2,564 corporate green bonds obtained from the Bloomberg Fixed Income database as of December 31, 2020, we examine the emissions implications of green bonds and its value for investors. We find three main results. First, the issuer concentration in this asset class influences the event study results. For instance, the cumulative abnormal equity market return (CAR) over three days surrounding the bond issuance event for listed firms in the international sample from 2013–18 is a modest 0.130 percent but statistically significant at 12 percent. However, the CARs are much higher for the U.S. sample (0.943 percent and significant at the 5 percent level) compared with the international sample (0.015 percent and insignificant). Further investigations revealed that roughly 75 percent of the U.S. green bonds were issued by Tesla or its subsidiaries. Remarkably perhaps, Tesla’s green bonds account for the bulk of the positive equity market reaction to U.S. green bond issuances. The CAR associated with Tesla’s green bond issuance is 1.136 percent (significant at the 5 percent level), while that associated with the green bonds for the remainder of the U.S. sample is insignificant. A similar issuer clustering in this asset class is observed in other parts of the global market. For instance, Vasakronan AB issued 29 percent of Sweden’s green bonds, Credit Agricole Corporate & Investment issued 59 percent of France’s green bonds. Similar patterns exist in other countries. Therefore, the market reaction to green bonds announcements mainly captures the stock market reaction to the announcements of a few major companies such as Tesla. Researchers that work on corporate green bonds might want to be mindful of the issuer clustering in green bonds.
Second, unlike in the primary market, where there is no yield differential for green bonds in the primary market, the secondary market of green bonds has a lower yield of negative 32 basis points relative to a propensity score-matched sample, a finding primarily attributable to the green bonds issued by the financial sector. However, for the polluting sectors of the economy (i.e., the utility, energy, material, and industrial segments), we found a negative 6.7 bps yield, suggesting that the green bonds of the main polluting sectors are perceived as riskier by the investors and therefore traded at a discount (relative to other green bonds). This pattern is somewhat surprising because higher greenium (green premium) values should be expected in the polluting sectors given the investors’ desire to provide financing to these companies to improve their environmental impact.
Third, we evaluated the value implications of green bonds for issuers. Analysis suggests that firms with a lower environment score in terms of ESG rating, higher carbon emissions, and a lack of carbon reduction target are more likely to issue green bonds as a sign that they intend to reduce their future emissions or improve their future environmental performance, however this may be defined. We found, though, that the emissions of green bond issuers do not fall even after four years following the issuance, a result that holds up even when focusing only on the emissions data voluntarily disclosed by the issuers rather than on those estimated by the data vendor. Hence, the environmental performance of green bond issuers suggests that the commitment to curb future emissions is both weak and unreliable.
Overall, our work raises questions about the value of green bonds for investors and the environment.
 For the sample spanning 2013–20, the proportion of U.S. green bonds attributable to Tesla or subsidiaries was 46 percent.
This post comes to us from Jitendra Aswani, a post-doctoral fellow at Harvard University and Shivaram Rajgopal, the Kester and Brynes professor at Columbia Business School. It is based on their recent paper, “Rethinking the Value and Emission Implications of Green Bonds,” available here.