Urban Returns

Sorting

I've written recently about Glaeser and Resseger's research showing that workers in skilled cities tend to become more productive as their cities grow while workers in unskilled cities do not.   I've also written about Abel et al.'s paper showing that while workers in all cities tend to become more productive as their cities grow denser, workers in skilled cities are especially likely to benefit.  The latter  receive, on average, three times the productivity gains of workers in the least skilled cities.  Growth and densification are good if yours is a highly skilled city.  

But the Glaeser and Abel papers also raise another possibility:  perhaps skilled cities are better off keeping out the unskilled than growing without bound.   

This is arguably an implication of their work. (And one I don't like, by the way.)   Both Glaeser and Abel define highly skilled cities to be those with high percentages of college-educated workers.  The higher the ratio of B.A.s to high school grads, the more skilled the city under their definition -- and the more the workers benefit as the city grows.

One way to keep a high ratio of skilled to unskilled is to price out the unskilled.  The most effective way to price unskilled workers out of a city is to keep the cost of housing high.  And, indeed, the highly-skilled cities on the coasts are adept at (and notorious for) using rigid land-use regulations to inflate home prices.  (Or were until recently; more on that later.)   The incumbent homeowners in these cities of course benefit from a tight housing supply since it raises the value of their properties.  But perhaps their skilled workers benefit too.

I speculated about this  last October without the benefit of Glaeser ir Abel's research:

Perhaps the elites in the superstar cities sense it is better to surround themselves with a uniformly high-quality workforce.  They want to spend their time with other elites; letting in lots of less-skilled workers would introduce so much static, just as if Harvard were to throw open its doors to anyone with a half-decent high school transcript.  In other words, an influx of less-skilled workers might dilute the experience for the high-skilled.

I got my undergraduate degree from Ole Miss and my law degree from Yale.   I studied harder at Ole Miss, took harder classes and spent a lot more time with my professors.  I thought law school was mostly boring and had a suburban home and the premium cable channels by my third year.  But I probably got a better education at Yale than Ole Miss merely by hanging around a bunch of people who were smarter than me.  There were economists, physicians, mathematicians, political scientists, a Broadway actor, an Olympic miler  and, of course, a bunch of smart new college graduates  from all over the country (but mostly from the Ivy League, Standford and Berkeley.).  It was a small class of 175.  It was impossible not to keep up with new ideas or trends because someone you bumped into would know about them.  I wasn't paying a gazillion dollars for the law school classes, I was paying for the spillovers.  It was a fair price.

The housing markets in  LA, San Francisco and New York function as giant sorting machines.  By setting home prices so high, they weed out all but the most skilled -- except for the class of relatively unskilled who are supported by price controls or subsidies.  They're using the  Ivy League model.

So are they better off?  Yes, clearly, if they can shunt the less skilled elsewhere and  continue to grow and densify.  But it's hard to do both.  High prices, in fact, require that the housing supply be kept tight  in order for prices to stay high.  Otherwise, the new supply of housing would drop home prices and the riff-raff (from their perspective) would come flooding back in.  But perhaps merely increasing the density of skilled workers will yield the productivity gains yielded by  growth and densification.  Their motto could be, "We get smarter, not bigger."

I think this is probably wrong, though.  What matters is being close to other people with the same skill set; that's where the knowledge and productivity spills occur.  A software developer benefits from having a lot of other smart software developers around; a musician benefits from having a lot of other good musicians around.  Adding more software developers or musicians who aren't quite as skilled shouldn't dilute this benefit because we city-dwellers largely control whom we interact with.  Firms, in fact, segregate themselves by quality all the time.  Deepening the labor pool simply can't hurt.  I think cities like San Francisco and New York would be even more productive if they loosened their growth restrictions and allowed in more people.

And even if were true that the San Francisco model can achieve greater productivity without growth, there's always the risk that its  sorting machine might break. To wit, a housing market collapse.  Skilled workers cash out what they can and move elsewhere.  Lower home prices attract the less-skilled who had been priced out of the market.  A political and regulatory climate makes the city less hospitable entrepreneurs, who benefit from cheap space and light regulation.

It will be interesting to see what happens to productivity in these wealthy, slow-growth cities as the recession wears on.  My guess is that they will have to begin growing vigorously again if they want to recapture the productivity they had at the peak of the housing boom.

Density and congestion, etc.

Whenever I see new data on cities, I'm always tempted to match them to the cities' weighted densities, if for no other reason than no one else does it.  And so with the Texas Transportation Institute's latest report on city congestion.  TTI found a wide range in hours lost due to congestion per year -- e.g., in 2007, Lost Angeles drivers lost an average of 70 hours per year to congestion;  Cleveland drivers, just 12.  (The relevant TTI table is here (pdf).)  Does weighted density partly explain this variation?

The answer is "No," based on my admittedly simplistic analysis.  

Below the jump I have three charts plotting, for 33 cities, weighted density, standard density and total population against hours lost per traveler to congestion.   (The 33 cities include the 31 largest urbanized areas -- excluding New York City, which is always an outlier -- and Austin and Honolulu.) 

As the scatter plot shows, weighted density explains virtually none of the variation in congestion (adjusted R² = .06).   Standard density explains a bit more  (adjusted R² = .19).  And total population, a bit more than standard density  (adjusted R² = .28). 

Skilled cities II

While I'm in this nerdy mood, let me point to this paper by Jaison Abel, Ishita Dey and Todd Gabe:

We estimate a model of urban productivity in which the agglomeration effect of density is enhanced by a metropolitan area’s stock of human capital. Using new measures of output per worker for U.S. metropolitan areas along with two measures of density that account for different aspects of the spatial distribution of population, we find that a doubling of density increases productivity by 10 to 20 percent. Consistent with theories of learning and knowledge spillovers in cities, we demonstrate that the elasticity of average labor productivity with respect to density increases with human capital. Metropolitan areas with a human capital stock that is one standard deviation below the mean level realize around half of the average productivity gain, while doubling density in metropolitan areas with a human capital stock that is one standard deviation above the mean level yields productivity benefits that are about 1.5 times larger than average.

This is a little technical, so let me translate:   Cities tend to become more productive as they grow denser.   On average, a city's workforce becomes 10% more productive when the city doubles in density.   But that average obscures the importance of skills.   Less skilled cities benefit a lot less than skilled cities from densification.   In fact, skilled cities, on average, enjoy three times the productivity gain from denser growth than less skilled cities.   This is yet more evidence of the increasing returns and agglomeration benefits from density.

This is a nice complement to the Glaeser and Resseger paper.  Glaeser and Resseger found that workers in skilled cities become more productive as the city grows, while workers in unskilled cities do not.   Abel et al find that workers in skilled cities become more productive as the city grows denser; workers in unskilled cities, less so.

The authors also test their conclusions using a variant of weighted density. They find even greater productivity gains (20% on average) when a city doubles its weighted density.

The authors use a coarse form of weighted density.  They weight urbanized area density by county subdivisions.  But how you chop up a city matters when calculating weighted density.  In order to calculate weighted density, you first divide the city into a bunch of smaller regions.  You then assign each region's density a weight equal to its share of the populations.  In general, weighted density increases as you chop up the city into smaller regions.

I used census tracts for my weighted densities.  There are many more census tracts than county subdivisions.  I thus got a lot more stratification than they did -- e.g., their top weighted density was 19,000 ppsm, while mine was 33,000.  How you divide a city for calculating weighted density is somewhat arbitrary, but I think using census tracts makes more sense than county subdivisions, which are more or less arbitrary.  I suspect the authors would have found even greater returns to density had they weighted density by census tracts rather than by county subdivisions.  

H/t Richard Florida.

Demand for suburbs vs demand for central cities

Americans overwhelmingly prefer suburbs.   We know this because most Americans  choose to settle in suburbs.

No, wait.  Demand for suburbs and exurbs is collapsing; they are on their way to becoming the next slums.

Both memes are wrong, I think.  The "suburbs are dying" meme is wrong because most metropolitan growth will continue to occur in the suburbs -- at least in rapidly growing metropolitan areas -- regardless of shifts in preferences.  The first is wrong because rapid suburban growth doesn't tell us anything about the relative demand for suburbs and central cities.  

First, rapid growth in a metropolitan area is synonymous with rapid growth in the suburbs.  And it will always be that way.  (The form of suburban development is a separate issue.)  It is harder to add new housing in a built-out environment.  Denser construction is more expensive and takes a lot longer to get built.  

Take Austin.  Austin's urban core had about 170,000 residents in 2000.  Austin's metropolitan area added about 390,000 residents between 2000 and 2008.  Even if Austin's urban core had doubled in population between 2000 and 2008, most growth had to and did occur at the metropolitan area's fringes.  Of course, Austin's urban core did not double in population.  It is not possible to double the density of a built-out urban core in just eight years.  I'd be surprised if Austin's urban core grew by 20% between 2000 and 2008.  So its suburbs necessarily have boomed, even though Austin's core is healthy and experiencing steady growth.

Ditto with cities like Dallas, Houston and Atlanta, which each added one million residents between 2000 and 2008.  It would have been impossible for their central cities to add more than a small fraction of the newcomers.

But because most growth must occur in the suburbs, this growth does not alone tell us whether demand for the suburbs is growing faster than demand for the central city.  

We know that demand is higher for the suburbs when the central city is losing population while the suburbs are adding population.  But in thriving cities, both suburbs and central cities are adding population.  The central city is simply adding population at a slower rate.

To figure out relative demand, we also have to consider prices.  Because the housing supply in the central city is relatively inelastic, higher demand mainly shows up as higher prices.  Because the housing supply in the suburbs is relatively elastic, higher demand mainly shows up as more subdivisions.  To determine whether demand is rising more rapidly for one than the other, we have to weigh both the rate of new construction and the change in prices.  And, to complicate things, we have to factor in shocks like rising gas prices and recessions; these have different impacts on the relative demand.  

So determining whether demand for one is growing faster than demand for another means disentangling a complicated knot.  That is a hard thing to do.  A lot harder than focusing on just rising supply or rising prices.

"Sprawl" \'sprol\ n.

The Overhead Wire asks what sprawl is.  Ryan Avent gives his answer -- he uses "sprawl" to refer to neighborhood density rather than regional density.  (As a stats geek, I would state it this way:  "What matters is a metropolitan area's weighted density and not its standard density.")

My own thought it that it's time to retire "sprawl."  This little 5-letter word is being forced to do too much work; it simply has too many meanings.  And it might as well be a 4-letter word because it is the land-use equivalent of "asshole" -- I've told you I don't like you but I won't bother getting specific. 

Here are different meanings I've seen it given: (1) peripheral development -- greenfield development at the fringe of an urban area (one variation singles out "leapfrog" development); (2) economic dispersal  -- the rising cost (in both time and money) of getting from point A to point B; (3) "pod" development -- any development sharply segregating single-family, multi-family and commercial uses; (4) low-density development  -- always a relative metric, of course; and (5) auto-dependent development.  I'm sure there are others.

"Sprawl" literally means "to spread or develop irregularly." That best matches (1). But what makes it such a slippery word is the correlation between these definitions, which causes people to conflate one meaning with another. Peripheral growth can lengthen commutes. Virtually all new suburban development is "pod" development, and low-density development at that. And everything built more than a quarter-mile from a rail station -- virtually any suburban development -- is car-dependent, or at least car-oriented.  It is therefore unsurprising that people use "sprawl" to encompass several meanings at once.

For example, Dolores Hayden in A Field Guide to Sprawl defines "sprawl" as "a process of large-scale real estate development resulting in low-density, scattered, discontinuous car-dependent construction, usually on the periphery of declining older suburbs and shrinking city centers."  So she combines (1), (4) and (5), and possibly (3) as well, and throws in a jab at tract development for good measure.  (Note, though, that combining meanings technically narrows the definition rather than expands it.)  There are many others, of course.

But it's sloppy to mix them up.  The correlation is imperfect.  New Urbanist developments that mix single-family, multi-family and commercial uses into a walkable neighborhood are often built on greenfields at the periphery of urban areas and are not served by transit.  Density is not a perfect proxy for urban form because pod developments are sometimes denser than traditional, walkable neighborhoods; this is one reason why places like Phoenix have a relatively high weighted density.  Infill development is frequently not served by transit and consists of segregated uses.  Commuting costs can rise with rising density as well as increasing geographic dispersal if the transportation infrastructure is not up to snuff.

Each of these definitions suggests a different problem.  You cannot propose a solution until you have told the rest of us what problem you are trying to solve.  And "sprawl" is simply too mushy to do that.  If you are worried about the loss of open space, then New Urbanist greenfield development must be included within the definition of "sprawl."  If you are worried about rising commuting costs, it makes no sense to include greenfield development served by mass transit.  If you care about urban form -- if you hate pod development, for example -- you should include pod development whether it's core or suburban, dense or sparse.  We're engaged in lots of different debates over land use but we often don't realize it because we use just one word to characterize them all.

We need to develop more precise, clinical terms to advance the discussion.  Medicine, math and science (and law, more or less) define their terms precisely.  People make fun of jargon, but serious disciplines need it to keep distinct ideas straight.  As used today, "sprawl" is the medical equivalent of "sick."  Imagine the hilarity that would ensue if doctors used "sick" as their primary term for both prostate cancer and the flu.  

Define terms precisely.  And don't build moral judgments into them.  You can cast aspersions once everyone else knows what you're talking about.

Potential demand for density

I noted recently that the 30 largest U.S. urbanized areas (excluding NYC) display no association between density and the number of households with children (as a percentage of all households).  I have put the numbers below the fold for the curious.

Interestingly, Tampa and Pittsburgh have the lowest percentages of households with children (I thought San Francisco would be at the bottom; it ranks just above of Tampa and Pittsburgh), while LA had the third-highest percentage, ahead of Dallas, San  Antonio and Atlanta.

The chart illustrates my point about density's irrelevance perhaps better than my scatterplot.  For example, the urban densities of Philadelphia, St. Louis, D.C., Detroit, Las Vegas and Cincinatti range from 3,274 ppsm to 8,457 ppsm, but their percentages of HWCs vary by just 4/10ths of one percent (between 35.3% and 35.7%).  (Remember, "urbanized area" includes the 'burbs contiguous to the urban core.)

The other thing that struck me was that households without children are a large majority in almost every large city.  They exceed 60% in all but five of the thirty.  Since I think such households tend to be more tolerant of density than households with children, there may be more market demand for infill density than most people think.  Neighborhood-activist types often complain that we don't need all those apartments and condos because families with children don't want to live in them.  Even if they're right about families with children, those households are a relatively small segment of the market.  

By the way, I am not suggesting that singles or couples would rather live in a 100-unit condo than a single-family home.  I imagine that, all else being equal, most singles or couples would rather have a single-family home.  My wife and I owned our own house for years before we had children.  And I certainly am not suggesting that singles or couples ought to live in multi-family units.  Not at all.  People should get the kind of housing they want.

But preferences depend on price. I prefer a 2009 Infiniti M45 to a twice-wrecked 2002 I35, but I drive the I35.  As home prices rise, and all else being equal, small households are more willing (on average) to switch to denser housing than larger households with children.  Their aggregate demand for SF housing is simply more elastic, more sensitive to price.  Rising property values really do generate more demand for dense housing, regardless of what families with children want.

Weighted density and households with children

There is a stereotype that dense places are filled with singles and childless couples while low-density, suburban cities are filled with families with children.  I'm sure this stereotype is true if one compares central cities to suburbs.  It's true for Austin, for example.  Children make up less than 15% of the population of San Francisco proper.*

But I was curious whether this stereotype holds when we compare apples to apples -- i.e., urbanized areas to urbanized areas.  Are denser urbanized areas less likely to have households with children?

I thought the answer would be "Yes," but it is "No."

The chart below plots the weighted densities of the 30 largest urbanized areas (excluding the outlier New York City) against the percentage of households with children.   As you can tell at a glance, there is essentially no association between the two:

The line is pretty darned flat.  A 1,000 ppsm increase in weighted density is associated with just a one-tenth of one percent decrease in the percentage of households with children.   (It is not a statistically significant association.)

This is counter-intuitive.  It's also strange given the fairly strong association between weighted density and renting:

 

In this case, a 1,000 ppsm increase in weighted density is associated with a 1.2% increase in the percentage of households that rent.  Weighted density accounts for about 30% of the variation in form of household tenure.  It is a statistically significant association.

This is strange because families with children, who comprise the majority of households with children, mostly own their homes.  Perhaps very dense pockets soak up households without children and leave more single-family homes for households with children.  In other words, density might change the distribution of households within the urbanized area but not affect the overall composition of households.  Dunno. 

This is a pretty crude analysis so don't take it too seriously.  But be skeptical the next time someone claims that densification runs off the children.  It might just move them around instead.

*Dear friends without children:  I am not exalting children or parenthood or denigrating those of you who are single (or those who rent).  Promise.  I'm simply curious about the effect of density on the urban environment, and the presence of children is a nice, easy variable to measure.  

Glaeser: Why New York City home prices won't collapse

Because it is dense:

I won’t be surprised if Manhattan prices do drop in the next few years, but I also strongly believe that the future of New York City continues to be bright. Homo sapiens are a social species; almost all of what we know we learn from each other. Dense cities, like New York, succeed when they take advantage of this fundamental aspect of our humanity. They thrive by enabling us to connect with each other, which then promotes learning and innovation. The current downturn will only increase the returns to being smart, and you get smart by hanging around smart people. As long as New York continues to attract and connect those people, the city will continue to thrive.

This supports a point some commenters made in response to one of my posts on the housing bubble, although Glaeser is crediting demand rather than tight supply.

Why do Southerners drive more?

As I noted in my last post, Kahn and Glaeser found that Southerners drive more than people in other parts of the country even after controlling for metropolitan area and census tract density.  I've been puzzling over that.  Here are some possible explanations:

1.  Southerners just like to drive more.  The South birthed NASCAR, after all.

2.  Southern cities tend to be less dense than cities in other parts of the country.  Kahn and Glaeser supposedly controlled for that.  But they assumed a linear association between density and driving -- i.e., an X% increase in MSA density leads to a Y% decrease in driving.  The relationship may not be linear.  Denser areas of course support better mass transit.  But density surely offers increasing returns to scale to mass transit.  Doubling residential density, for example, supports a denser network of bus routes, which makes buses more attractive and increases the probability that a given resident will ride the bus.  Put differently, doubling residential density might increase ridership by 120% rather than just 100%.

3.  Southern cities tend to be poorer, which means they have less money to spend on buses.  (Of course, this also means their residents have less money to spend on cars.  But of course their housing is cheaper so their residents have more money to spend on cars.  And so on and so on.)  

4.  It's a legacy of racial discrimination.  When I lived in Jackson, Mississippi, I noticed that bus riders were overwhelmingly black; I assume that whites abandoned buses once the ridership mix tipped decisively toward blacks.  (The same dynamic operates in public schools).  This likely has three consequences:  (1) Most obviously, almost all whites drive; (2) because almost all whites drive, there is poor bus service in predominantly white neighborhoods, which reinforces whites' disinclination to ride the bus; (3) because bus service is mainly used by blacks, whites aren't inclined to vote much money for it.

I'm sure I'll come up with others five minutes after I hit the "Publish Now" button.

Density and returns to education

The Glaeser/Redlick paper I discussed yesterday really is worth a read.  It makes two very interesting points about the benefits of density:

The returns to skill are also higher in denser areas.  As density doubles, skilled workers earn 5,400 dollars more relative to the less well-educated.  This effect is quite statistically significant.  One interpretation of this fact is that density and skills are complements; this is also supported by the propensity of high human capital industries to locate at the center of urban areas.

And this:

[W]e show that there is also a significant interaction between area density and individual skills.  As density doubles, the college premium appears to rise by about 4.6 percent.

Neither result necessarily means that density itself increases the returns to skills and education.  Density is correlated with high property values, and property values are high when a location offers high returns to labor.   But the correlation between density and the returns to skill may also reflect the increasing returns from positive spillovers; workers tend to be more productive when they are close to one another.

In either case, restricting density negatively impacts welfare.  The restrictions either they shut some people out of high-return locations, or they dampen the increasing returns to scale offered by a denser environment.

I do wonder, though, whether these statistical associations would hold if run using weighted density rather than standard density.  Hmm.