Virtually without exception, geographic models of network investment and optimization have ignored the implications of the methods of taxation and finance used to generate new transport facilities. It is standard to cast these problems in a joint fixed and variable cost framework or to minimize variable cost, subject to an exogenously specified budget on fixed costs. To be sure, this simplistic view of financing arrangements in the transport sector can be defended on the basis that more serious deficiencies in these models are of greater concern-for example, the failure to account adequately for transport system economic interdependencies. However, since the taxation and financing arrangements themselves affect the spatial pattern of road impact, it seems timely to examine their implications in a dynamic geographical road model. We take a step in this direction by combining a quite simple geometric network representation with a fully dynamic economic model and an explicit taxation and road financing framework.Specifically, this paper concerns itself with a comparison of three different road financing schemes that could be used by an authority faced with the problem of defining the optimal road penetration schedule from a coastal port into its hinterland. The effects of road extension are captured in a fully dynamic, recursively structured, single-commodity market centered at the port. The three taxation schemes include:(1) a tax on incremental rent (partial land value capture), (2) a ton-mile or vehicle mile tax, and (3) a commodity sales tax. We consider each method as a means of financing road extension and show that each method gives rise to a different pattern of spatial development around the port, to a different level of benefits, and to a different equilibrium condition.The structure of the paper is straightforward. First, we outline the road development model and relate it functionally to the dynamics of the commodity market. Second, we identify the costs and benefits of road penetration and specify the general form of the road extension problem. Benefits are taken to be only those occurring to producers. Other possible benefits, such as agglomerative economies at the port, are ignored. Next, we describe the impacts of the three taxes and derive the revenues from each method. Then, the three taxation methods are embedded *Preliminary work on this topic was supported by the National Science Foundation Grant SES Bruce A. Ralston is associate professor of geography, University of Tennessee. Gerald M .
8105838.Barber is associate professor of geography, University of Victoria, British Columbia.