This study presents a general methodology for detecting indirect per unit costs of multi-product firms. The methodology rests on the law of "no distortion" of the decision making process invoked by SCAPENS (1991). The game theory approach (SHUBIK (1962), HAMLEN, HAMLEN, TSCHIRHART (1980)) has previously exploited the idea of neutrality of joint cost allocation. It constrains the allocated costs to be conditional to the decision making process of the firm (DEMSKI, FELTHAM (1976)). We admit here that the aim is the maximization of the profit. We use the law of "no distortion" to derive a cost allocation procedure that is neutral. It says that under a fixed decision objective the optimal action is the same whatever the joint cost is allocated or not. When the main objective of the firm is to optimize its profit, the condition of "no distorsion" of the profit maximization processis the following one. The proportionality of the allocated costs and the proportionality fot the marginal costs must be the same for all pairs of products. The only cost allocation method that replies on this understanding presents as a procedure of distribution in proportion to the cost volume elasticities. This procedure involves that allocated costs reflect marginal costs. Moreover, it explicitly takes into account the impact of economies of scale. We show that the dependence between allocated costs, marginal costs and economies of scale, is intrinsic to the problem of the joint cost allocation. Therefore, this last one becomes truly relevant only when the joint cost fonction is not homogenous. The law of "no distortion" is very general since the rusulting neutral allocation rule is closely dependent upon the assumptions concerning the profit function. Thus, the methodology extends without difficulties to more sophisticated cases. For example, we could use the information contained in the complexity of the production process (may be expressed in terms of activity centers). One also could extend the analysis to the case where the economic environment is uncertain, risky and intertemporal