So here’s something. Way back in January, Tyler Cowen recognized a new blog at the NY Sun covering the topic of congestion. I had a look and found it wanting. For some reason, the Sun blog chose today as the day to venture a response. Here, let me just post this:

A blogger, Ryan Avent, rants about my original discussion on population density in LA versus New York. Suffice to say, he should read the follow-up posting on density, in which Wendell Cox explains that his data come directly from the Census Bureau.But Mr. Avent goes on to question my economics creds and my familiarity with the concept of “externality” as it relates to NYC transit:

There are plenty of relevant ones here to choose from — carbon emissions, congestion, the spillover effects of density in NYC that help make the city an economic juggernaut — all of which indicate that the state should, in fact, be subsidizing transit.

To me these sound very much like reasons — “negative externalities,” since he insists on using economics lingo — for the state NOT to be subsidizing transit. Anyway, is he saying that absent the subsidies NYC would not be “an economic juggernaut,” or that being an economic juggernaut, ipso facto, justifies subsidies? The first is far-fetched, the second a non sequitur.

The first point pertains to the suggestion that there is anything meaningful about the statistic that Greater Los Angeles has a higher population density than Greater New York. I wondered about Mr Wendell Cox’s statistics, it’s true, because it appeared that he included or excluded the Riverside MSA depending on how it affected his numbers. In the followup numbers he sends to the Sun, he specifically does not include Riverside. This is pretty misleading, since Riverside is clearly a part of the LA economic sphere and also happens to be ultra low-density.

But as I wrote before, and as Sandy Ikeda completely ignored, this doesn’t particularly matter. Average density is a foolish measure to use here. The New York core is far, far denser than Los Angeles’, and the only reason LA comes out ahead is because its suburbs are geographically constrained by topography and are therefore denser than New York’s distant exurbs. No one in their right mind would visit New York City and come away thinking it was less dense than LA. That’s because it isn’t, not in any meaningful sense.

I went on to wonder whether the Sun blogger understood the concept of externalities. The answer is clearly a resounding no. Transit displays positive externalities because riders using the system benefit others not involved with the rider’s decision to take transit. By taking the train, a rider reduces highway congestion. He also reduces (substantially) his per capita carbon emissions. And the economics literature demonstrates that density is directly related to productivity. Without a comprehensive transit system, Manhattan could not fit so many jobs in such a small geographic location–office capacity would either be crowded out by new residential capacity, or by the infrastructure needed to support a massive daily influx of automobiles. As a result, productivity and wages in Manhattan would fall.

When there are positive externalities to goods, they are underprovided by the market and should be subsidized by the government. There are quite reasonable analyses arguing that transit subsidies up to 90 percent are justified.

So there you are, Sandy. Let me know if you have any additional questions. I expect to hear back by, I don’t know, June?


  1. AC says:

    A better metric, one that probably does a better job of capturing the difference between LA and NY, is this:

    Rather than divide total population by total land area, compute a weighted average using census tract. Multiply the total population of each census tract by its land area. Sum all of these and then divide by the sum of the areas of all the census tracts.

    This weighted density discounts large, sparsely populated census tracts, and gives extra weight to densely populated tracts. You could call it “average perceived density.”

    In a city with a more or less uniform density, weighted density might equal standard density. But in a City like NY, weighted density will be much higher than standard density, because the bulk of the population is concentrated in a small area.

  2. ryan says:

    Exactly. Question: has someone computed weighted average densities? I’d love to see them.

  3. AC says:

    I haven’t seen weighted densities computed. I might take a shot at NY’s sometime using zip codes.

    (I’ve been blogging long enough to read what I write before posting. The formula for weighted density is: (1/P)* Sum((each tract’s population squared)/each tract’s land area),
    where P = total population of all tracts.

    Basically, it’s each tract’s density times that tract’s percentage of total population.

    Run it with two tracts, one 10 sq. mile with pop density of 100/sq mile and one 20 sq miles with pop density of 5/sq mile. The “standard” density is 37/sq mile, while the weighted density is 913/sq mi.)

  4. BeyondDC says:

    Anything attached to the names Wendell Cox, Randall O’Toole, or any number of conservative-sounding “institutes”, “foundations” and “centers” (all run by Cox and O’Toole) should be immediately dismissed as intentionally biased, deliberately anti-urban and anti-rail, and underwritten by GM.

    Serious people stopped listening to the likes of them ages ago. Nothing they say can be taken the slightest bit seriously.

  5. AC says:

    Ryan, I’ve calculated “weighted averages” for NY and LA using census data. The census department breaks down a metropolitan area into various “central places,” and gives the population and density for each. It lumps everyone else into the “not in central place category,” and gives the density for this population.

    Using this data, I calculated weighted average density using the formula described in my second comment. I got:

    LA: 7,960
    NY: 14,786

    I.e., NY is almost twice as dense under this method, something we all intuitively know.

    By comparison, the Census Department lists these as the average densities for the metropolitan area:

    LA: 7,068
    NY: 5,309

    Note that LA’s weighted density is not much greater than it’s MSA-wide average density.
    I’m preparing a post on my blog with data for different cities, along with a more detailed description and justification of this methodology.