We discuss whether Basu's (1986) model of triadic power relations is robust to generalizations where we allow multiple landlords, merchants and laborers. In the special case of one landlord and multiple merchants we show that the landlord's threat towards any laborer becomes credible even in the original stage game model. For the case of multiple landlords we need to generalize more recent solution concepts. We add realism by allowing the laborers' reservation utilities to vary with the costs of trading in external markets and the characteristics of the laborers. We allow production by landlords and merchants, as well as Cournot competition among merchants. In equilibrium the rural wages are a function of the number of landlords and merchants, the characteristics of the laborers and distance to external markets. We estimate the model, using household survey data from Nepal, and find strong support for the triadic model.