A Master Face describes a, in most cases, semi-synthetic face image that achieves a significantly higher chance of false matches when attacking random face galleries. The methodology for creating a master face can vary. In this work we investigated if it is possible to generate such a facial image by current morphing techniques. Subsequently, we generated averaged faces within demographic morphing trees to analyze whether these averaged faces increase the probability of success on a false match in a face recognition system attack and whether there are any concrete differences for specific demographic groups. Two state-of-the-art face recognition systems were evaluated by attacking the demographic face galleries using the averaged faces. For the generation of these faces and the classification of the deposited face galleries, subsets of publicly available face databases were created and classified. To analyze the identification systems, only the rank-1 comparison score was evaluated in the experiments for each averaged face. For the majority of experiments, the tendency of increased false matches decreases as the attacking faces become more average. Only the Asian demographic groups showed a strong deviation from these trends, scoring significantly higher on average than the other groups even for averaged faces from deeper morphing trees.