While there are growing bodies of research examining both the differences between strongly and poorly socially performing firms, and the impact of firm leaders on other strategic outcomes, little has been done in examining the effect of firm leaders on corporate social performance (CSP). This study directly addresses this issue by using upper echelon theory, and the KLD Research Analytics CSP ratings, to show that observable CEO characteristics predict differences in CSP between firms, even when firm and industry characteristics are (...) controlled for. Using a sample of 650 public US firms I find that strong or exemplary CSP, as measured by the strengths categories of KLD's ratings, is positively related to the CEO having a bachelor's degree in humanities, having a breadth of career experience and being female. I find that KLD strength ratings are negatively related to the CEO having a bachelor's degree in economics and to their level of short-term compensation. Preliminary tests of causality support the assertion that these effects reflect CEO discretion rather than being an artifact of reverse causality. Significant relationships between the CEO characteristics and poor social performance as measured by the concerns categories of KLD's ratings are not found. This suggests that CEOs have more discretion in influencing strong and exemplary social performance than in impacting poor CSP. Implications, particularly for economics education, are discussed. While there are many fruitful directions for future research, the benefits and challenges to conducting similar studies in other countries is focused upon to correspond to the global nature of this special issue. (shrink)