Social bonds and the "social premium"

Social bonds (SB) have witnessed an unprecedented increase especially since the outburst of the Covid-19 pandemic, but their performance vs. conventional bonds (CB) has not yet attracted attention in the academic literature. As far as we know, this is the first paper to test the existence, the sign and the determinants of a "social premium", which we propose here to define as the yield differential between a SB and an otherwise identical CB. To this end we set up a sample of 64 SB aligned with the International Capital Market Association principles and 64 (exactly) matched CB, from October 2020 to October 2021 so as to focus on the peak of SB issuances. Regressions are based on the hypothesis that daily yield differentials between SB and CB may be determined by differences in non-perfectly matched characteristics. Based on the FE specification, which turns out to be preferred vs. OLS and RE both theoretically and empirically, two main results emerge. First, the social premium is significantly explained by differences in liquidity and in volatility, which are, respectively, negatively and positively correlated with the yield differential. Second, on the whole sample, the analysis of the fixed effects proves the existence of a significant and positive social premium that amounts to 1.242 bps. This result is robust to outliers, but differences emerge on subsamples especially in relation to issuer sector, thus pointing to the relevance of the use of proceeds, an issue that deserves further investigation as the SB market becomes more mature.