This paper studies majority voting outcomes for a specific class of two-dimensional policies. One policy instrument influences efficiency and the other redistribution. Absent the political process, the two dimensions can be addressed separately. With a two dimensional vote, the two aspects will interact in a non-trivial way. The illustrative policy we consider, requires taxing an externality-generating good and determining a budgetary rule which specifies the proportions of the tax proceeds that go to wage earners and to capital owners. We show: First, a sequential vote wherein the tax rate is determined first and the budgetary rule second, always possesses an equilibrium and that this equilibrium is the median-endowed individual's most-preferred policy. Second, the reverse sequential choice implies that the median-endowed individual may, but need not, be decisive. Third, the ``Shepsle procedure'' also implies that the equilibrium is the policy most favored by the median individual. Fourth, this equilibrium constitutes, under certain circumstances, the Condorcet winner for the unrestricted simultaneous voting game.
