This paper is concerned with the attempt to model and simulate an urban housing system dominated by non-market social housing, primarily to forecast demand for social housing under different scenarios. The urban system concerned is the city of Glasgow and its suburbs, a post-industrial city in West Central Scotland, a region now emerging from long-term structural economic decline. There is an established literature concerned with the development of metropolitan housing market models in both the USA and the UK. The present model draws from these traditions but is heavily influenced by the work of Meen (1999). The Glasgow model is heavily demand-determined with only a limited supply-side but with a standard market-clearing setup. Data for the model comes from the Scottish House Condition Survey 1996 and from extraneous housing, population and household estimates from local authority planners. The focus of the core part of the paper is primarily on the demand-side. Demand in the model is composed of three elements: new household formation, net migration and the tenure and locational choices of existing households. It is this third element that poses the most difficulties and is modelled separately using a nested multinomial logit formulation. The paper discusses the modelling issues and results from a series of NMNL models that attempt to explain the locational, tenure and mobility decisions of existing households. The preferred results are then adopted as conditional probabilities in the simulation model. The paper sets out the structure of the basic model and reports some initial runs. The paper concludes by examining the academic and policy implications of the model and suggests future avenues for refinement and further work.