Why Do the Poor Live in Cities?

The title of this post is borrowed from a 2000 NBER working paper by Glaeser, Kahn and Rappaport. Starting with the assertion, “that the concentration of poverty in central cities occurs mainly because such cities attract poor people, not because central cities make people poor,” (p. 1) it asks why.

Now that‘s a planning research question — a substantive, meaningful, attention-getting, take home to mom interrogative aimed at explaining cities. Much urban planning is about addressing the consequences of poverty and then, by extension, its causes. One part of the latter is why people are poor, another is where — lots of sorting questions in this blog — and it is easy to imagine that a good answer to one might also inform the other. So this is the kind of query you pursue when you really want to get to the bottom of things and also make a splash en route. Unless the answer is obvious or you don’t have a good plan for getting to an answer.

Obvious?
Making it not so are the many competing explanations. A prominent textbook version comes from Alonso/Muth model of the city as a place where commuters trade off access to work against shelter costs. It implies that the rich will want to live where land is cheap (the burbs) if the demand for land rises faster with income than transport costs. Very elegant but, as Glaeser, Kahn and Rappaport point out, the land vs. travel cost demand assumption doesn’t seem to hold up empirically.

Many also recognize that cities and suburbs differ in lots of ways, such as by various amenities (school quality, crime) and discriminatory practices, any of which could serve to sort by income or correlates of income. Apart from discrimination, however, Glaeser, Kahn and Rappaport argue that most such amenities are less a cause of poverty or its concentration as much as a consequence. In sum, “Traditional housing market explanations cannot explain the sorting of the poor into cities.” (p. 29)

Research plan?
One urban feature they do suspect plays a substantial role in the underlying attraction of the poor to cities is transportation, and particularly the availability of public transit. The principle appeal of transit, as a way of getting around, are its low cash fixed costs (no need to buy a car). Of course, everyone likes low costs but the rest of the package are very high marginal time costs, the value of which rise with income. The bulk of the paper tries to nail this down empirically, given that the availability of transit in cities (and not suburbs) is also a policy variable to be explained. (Transit is likely also provided, in part, to serve low income users. It may be both cause and effect, which has implications for the statistical strategy.) In the end they find evidence of a very strong, independent sorting effect for the role of transit. More than expected, they say.

A second “underlying force attracting the poor to cities are more redistributive government policies,” such as housing subsidies, social services, and other transfers. They also find a sorting effect there. The paper concludes by echoing the arguments of Kain and his colleagues in the 1960s that transportation, especially public transit, is an important policy tool for influencing income sorting across space (as are jurisdictional boundaries associated with differential redistributional policies).

The demand for sharing
A UCLA Ph.D. student, Kimiko Shiki, has a more complete explanation, inspired in part by her relative poverty as a graduate student. The simple version is that the poor “share” more than the rich. An example is relying on roommates to split housing costs. Others include crowding a bus or queuing for a washing machine at the laundromat. These are examples of greater reliance on public goods as income falls, or perhaps of a greater propensity to turn pure private goods into shared goods by increasing densities. Indeed, the concentration of povery is in part a “benefits of density & proximity” story. In a way, Shiki’s story is a generalization of the Glaeser, Kahn, Rappaport theory to any goods or services where users can choose to reduce the fixed cash cost by sharing, with an added potential scale effect.

Formally, she plans to base her conceptual model on the production externalities literature, though it seems to apply both to situations where proximity or sharing have positive externalities and where they simply lower average fixed costs. (I think the externality argument, and maybe the shared costs one too, could probably be restated in terms of social capital — or vice versa.) In either case, it is kind of an agglomeration economy story. I won’t give away her research design and data strategy.

p.s. The idea behind these posts is to summarize and comment on selected parts of research issues I find interesting, in a brazenly incomplete manner that nonetheless aspires to move the discussion forward. Think of them as asides. Comments and corrections are welcome.