Energy efficiency commitments, often associated with tradable energy efficiency certificates, dubbed "white certificates", were recently implemented in the United Kingdom and Italy and will soon start in France. Energy suppliers have to fund a given quantity of energy efficiency measures, or to buy "white certificates" from other suppliers who exceed their target. We develop a partial equilibrium model to compare white certificates to other policy instruments for energy efficiency, i.e., taxes and standards. Our conclusions are: First, if white certificates are chosen, each supplier's target should be set as a percentage of the energy they sell during the commitment period rather than in absolute terms, e.g. based on past variables. Indeed the latter solution decreases sharply energy suppliers' profit since they cannot pass the cost of certificate generation on to consumers. Such a system thus risks generating a fierce opposition from these industries. Furthermore, setting individual targets independently of the evolution of market shares seems unfair. At last, this system risks creating a large rebound effect, i.e., a large increase in energy services consumption. Second, compared to taxes and standards, white certificates (with targets in percentage of energy sold) seem particularly interesting to reach a certain level of energy savings while limiting distributional effects, thus to limit oppositions to its implementation. Furthermore, they generate less rebound effect than standards and seem more able than taxes to mobilise a part of the no regret potential. However if targets are too weak there is a real risk that white certificates systems fund mostly business-as-usual energy efficiency activities, thus having little impact while delaying the implementation of other policy instruments.