The dependency method was an original method to estimating the economic impact. It was applied to the horse industry in four French departments (Ain, Allier, Maine-et-Loire, Bouches-du-Rhône). The issues are both theoretical (estimating short and long-term effects by integrating Keynesian theories and transaction costs theory) and empirical (developing a flexible method at a territorial level and for a sector subject to various shocks). Surveys were carried out with a sample of companies to estimate their jobs and their dependence on the equine sector. An overall dependency score is estimated from three components: short-term dependency or relative importance of the horse industry in outlets; medium and long-term dependence, reflecting the company’s ability to adapt to changes in the horse industry; geographical dependence, which corresponds to the territorial constraints suffered by the company within the department studied. Results highlights that the dependency method gives estimations comparable to more traditional methods in regional economics (as the economic basis); but with an order of magnitude higher due to the consideration of long-term effects. Two sets of explanations can be put forward to explain the higher level of estimates by the dependence method. On the one hand, the theoretical foundations of the 2 methods are different: while dependency is a "dynamic" method (which captures the effects of transformation of the productive system), the classic models of regional economy (economic base, input-output) quantify only short-term Keynesian income effects. On the other hand, there are reasons related to the scope of activities: only professionals of the sector for economic base models; consideration of the amateur sector in the surveys implemented by the dependency method.