According to the model of HOEL (1991), a unilateral emissions abatement of a global pollutant leads to lower aggregated emissions in a game with a simultaneous decision protocol. Our experiment tests the Hoel model and examines the question of whether a leader can induce additional abatement of followers in a game with a mixed sequential-simultaneous decision protocol. Using an environment with a unique interior equilibrium, our experiments confirm the stylized outcomes of previous public good experiments. Changes in abatement and profits for the simultaneous decision protocol are in line with the theoretical predictions of the Hoel model albeit not significantly in every case. In the treatments with a mixed sequential-simultaneous decision protocol, during the first periods an abatement near social optimum is mostly chosen by the leader. However, in most cases the leader failed to induce cooperation, i.e. there are few followers who react cooperatively to the leader's signal. High efforts by the leader and the cooperative followers are exploited by the majority of defective followers.