Ed Glaeser writes:
The late Senator Daniel Patrick Moynihan of New York is often credited with saying that the way to create a great city is to “create a great university and wait 200 years,” and the body of evidence on the role that universities play in generating urban growth continues to grow.
I’ve always thought this too, that it’s too bad that, given the total cost, a lot more cities would’ve benefited, over the years, by maintaining great universities rather than building expensive freeways, RenCens, and so forth.
But Joseph Delaney argues the opposite, considering the case of New Haven, home of what is arguably the second-best university in the country (I assume Glaeser would agree with me on this one):
According to wikipedia, the poverty rate in New Haven is 24%, which compares unfavorably with the rest of the United States where it is 14%. The poverty rate in New Haven, despite the presence of Yale, is nearly twice that of the United States as a whole.
Now, one might note that many of the poor residents of New Haven are likely to be students. This is true. But these students still use municipal services and thus require the local tax base to support them (in addition to the long term residents). They do not (after they graduate and make additional income) send money back to New Haven so, in a sense, New Haven is actually subsiding the urban communities that Yale graduates move to.
So, it is actually possible that a large university in a small community could be a drag on the economy due to the lower per capita tax base. Plus, you have a large segment of the population with only a short term interest in the community which may make long term planning more difficult. And New Haven, CT is not the only university town that I can think of with high levels of poverty.
Argument by anecdote is problematic, I know. But Delaney makes some compelling points that I’d think could generalize. (And his coblogger Mark Palko notes out that Detroit is another example of a declining city near a great university.)
On the other hand, I could imagine that Glaeser has a quantitative analysis in which New Haven and Detroit are two atypical special cases, I’d be interested in seeing more.
More general, I think a key step in statistics is making the connection between individual cases and overall patterns. It’s true that anecdotes can be misleading, but I also believe that a model is only believable to the extent that we can understand where the individual data points fit in.
P.S. Further discussion in the comments here. Let me emphasize that I’m not trying to say that New Haven or Detroit shoot down Glaeser’s story; rather, I’m saying that it should be possible to understand any model through its special cases (which, after all, come together to form the general pattern). This is the discussion I’m trying to start here.
