Suppose that you are Paul Krugman. You are a benevolent planner who wants to maximize the well being of all 300 million people in the United States. Where should these people live and at what population density? Would you build 100 3 million person cities with the density of Houston or of New York City's Downtown? Would you build 10 Mega Cities or would you distribute the population uniformly across space?
Economists are always interested in how close does the outcome we see in our market economy approximate the "ideal outcome" that the benevolent all knowing planner could achieve.
To solve the planner's problem you would have to grapple with the question of the costs and benefits of cities and how these private benefits and social costs change as a function of city size. Ed Glaeser's 1998 Journal of Economic Perspectives piece ("Are Cities Dying?") offers a clear overview of the economic forces encouraging and discouraging the formation of big cities. Vern Henderson also posts several useful urban papers here:
If you would like to see my new paper: "Quality of Life in Sprawled and Compact Cities" a .pdf is available here: