What are the Prospects for a Politically Intelligent Planning System?

In “Prospects for an Intelligent Planning System,” Kitchin, Dawkins, and Young (this issue) describe a tantalizing vision for the modern urban planner – complete, longitudinal and spatial urban data in a Geographic Information System with enough computational power to generate predicted outcomes for regulations, investments, and development. I’ve spent my research and practice career thinking about how neighborhood housing markets change in response to planning and policy interventions, and have often wished to be able to demonstrate exactly how a city’s choices about urban regeneration will play out over space and time, and to forewarn planners about unintended consequences like displacement. Looking backwards, it’s easy to analyze the effects of our choices – but we plan for the future, and wish for our data systems to support decisions. My colleagues rightly list problems with data access and quality and the politics and institutions of planning as barriers to intelligent planning systems. In this short paper, I will describe how these challenges play out in my own research and practice consultancy work in Portland, Oregon. I attempt to take a data-driven approach to gentrification and housing displacement, trying to predict the effects of new bus rapid transit and light rail lines in racially diverse, lower-come neighborhoods. Will new service support communities have greater mobility, or will they lose out when they are displaced by higher-income residents as their neighborhoods gain access? My analyses focus on whether owners raise rents, convert units into higher-end housing, or sell to owner-occupants as transit changes the desirability of neighborhoods, and where low-income residents might move if they are displaced. Planning agencies in our region have significant modeling capacities and are known for coordinating land use and transportation systems data and innovating in ‘big data’ planning. By contrast, I have found significant barriers in availability and access to housing data, and methodological issues with the use of housing data in urban systems models. These problems with our ‘smart cities’ planning consistently and systematically undercount or miscount the people who are most vulnerable to urban change – low-income households, people of color, and renters. Obviously the first step towards an intelligent planning system for understanding the housing market implications of policy, is a dataset with a count of units, their price and quality. However, data that comprehensively covers housing, especially rental housing, are not available in the Portland region. That there are significant missing components of these datasets is not due to the capacity to collect or store data, but due to political and market features of a system that relies on private actors to provide housing