Austin Contrarian

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.

I'm twittering

Hesitantly.  I'm not sure I want people to know how dull I am, but they've probably figured it out already.

Abolish the U.S. Department of Transportation

Urbanists everywhere are furious (and rightly so) at the Senate's decision to slice mass-transit funding from the stimulus bill.  If nothing else, the federal government ought to help transit agencies keep their buses and trains running.  A recession is a bad time for service cuts. 

Ed Glaeser, a Harvard economist whose work I often cite, argues that capital spending on infrastructure is too important to leave to a stimulus bill, though.  Passing off capital expenditures as stimulus enabled Congress (again) to short-change cities:

President Obama is the first urbanite in the White House since Teddy Roosevelt. He certainly knows the vital role that cities play in America. Yet despite the Chicagoan on Pennsylvania Avenue, infrastructure spending in the House stimulus bill follows a business-as-usual pattern that discriminates against density. The only way to break that pattern is to take non-repair-related infrastructure spending out of the stimulus, and craft a separate bill that looks beyond the current recession. Major infrastructure projects, especially in cities, cannot be done quickly.

Per capita transportation spending in the House stimulus package, including transit, is more than 50 percent higher in the 10 least dense states than in the 10 densest states, including Massachusetts. Yet America's highways and rails already make it easy to move goods and people across America's open spaces. The hard slog is getting across dense downtowns. Other elements in the stimulus package also favor farm over city.

Glaeser ignores the core problem, though:  our federal system gives less-populated states too much influence over federal transportation spending.  Which means that out largest states -- and thus our largest cities -- have too little influence.    

This will never, ever change.  The only solution is to get the federal government out of the transportation business.  Abolish the U.S. Department of Transportation, except for the FAA and NTSB, and let metropolitan areas keep the transportation dollars they now send to Washington.

How could anyone who cares about cities oppose this? 

• Cities would no longer have to beg bureaucrats at FHWA or FTA for their fair share.  Cities could spend their money on the projects they believe would yield the highest returns.  Locals know their needs better than political appointees sitting in offices hundreds or thousands of miles away. Our transportation dollars would chase better projects.
• It makes no sense to redistribute transportation dollars from cities to the countryside.  Cities generate most economic activity.  City workers are more productive, on average, than rural workers.  Starving cities of infrastructure limits their growth, saps their vitality and encourages people to live in less productive areas.     
• A dollar spent on city infrastructure yields higher returns than a dollar spent on a rural road.  For example, a dollar spent maintaining South Lamar in Austin benefits tens of thousands of drivers a day; a dollar spent maintaining RR 967 in rural Hayes County, perhaps a few hundred.  This is because cities enjoy economies of scale -- they do not have to double their road network when they double in size.  By redistributing money from cities to rural areas, the feds move money from more productive uses to less productive uses.
• Letting cities spend their own money would incentivize them to develop efficient infrastructure.  Under the current system, cities angle for highway dollars when the feds are handing out highway dollars; they angle for transit dollars when the feds are handing out transit dollars.  Since cities choose the projects that maximize their share of the federal dollars, they do not have an incentive to pick the most efficient mix.  Dense cities, for example, need less infrastructure per capita than less dense cities.  I don't believe we should try to turn Houston into San Francisco; variety is important.  But cities ought to bear the full costs and benefits of their infrastructure investments. 
• Finally, DOT spends billions simply deciding how to spend its money.  Spending money to move money from more productive projects to less projects is just waste.  And then there are the billions of dollars spent by rent-seeking corporation whose profitabililty depends on DOT policy.

The United States Department of Transportation was created in the 1960s to coordinate development of the interstate highway system. Perhaps we needed a federal agency to do that. But the interstate has been built.  DOT now does more harm than good.

I realize I've ignored the enormous influence of state transportation departments.  Their influence is probably even more pernicious than DOT's. (I'm not very fond of TxDOT, for example.)  But let's take it one step at a time.

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.  

A high-speed rail proposal that makes no sense

There were serious proposals to build high-speed rail in Texas in the mid-1990s.  The proposals never made it out of the station, though, largely because of fierce opposition by Southwest and the other airlines.

There is a legitimate debate whether high-speed rail makes sense for Texas.  But this new proposal is any easy call because it is guaranteed to fail:    

The idea of high-speed rail is being pushed again in a big way in Texas, and backers hope to have $12 billion to $18 billion high-speed trains running by 2020. This time, they say they have taken care to ensure the idea won’t fall flat the way a bullet-train push did some 15 years ago.

In the past, high-speed rail was not completed in Texas primarily because it was a top-down model driven by lobbyists out of Austin,” former Harris County Judge Robert Eckels, chairman of the nonprofit Texas High Speed Rail and Transportation Corp., told lawmakers at a Wednesday transportation briefing.

This time, he said backers from the consortium — which includes elected leaders, cities, counties and two airlines among others — reached out to past opponents to try to solve their concerns. Among them: Southwest Airlines, which fought the last high-speed rail project as a potential competitor. Southwest spokesman Chris Mainz said the airline is neutral on this proposal.

The high-speed trains — with an average speed of 200 mph — would run to airports, allowing rail to work in conjunction with airlines by ferrying in passengers catching longer flights.

The rail would run along the so-called “Texas T-Bone” — from Dallas-Fort Worth through Austin to San Antonio, and branching off in Temple to Houston. More than 70 percent of Texans live in the area that would be served.

Texas airlines already offer frequent service among the "big four" -- Houston, Dallas, Austin and San Antonio (except for Austin to San Antonio). In order for high-speed rail to make economic sense, it must offer an improvement over intrastate airline service.  Merely being competitive won't justify the billions in construction costs.

I think high-speed rail travel might be better than intrastate flights if done right. To be better, the train trip (1) must not take too much longer than an airline flight and (2) must not be much more expensive.  I think (1) is likely, at least for downtown-to-downtown travel.  For me to travel to Dallas by plane, I must make a 20-minute drive to the airport; arrive at least 30- to 40-minutes before my flight to get through security; wait for the flight (and possible delays); endure a one-hour flight; and spend 20 minutes getting from Love to downtown.  In practice, I can rarely make the trip in less than two-and-a-half to three hours.

A 200-mph, high-speed train trip might take, say, three and a half hours door to door if we allow three hours for the trip itself (an 80 mph average speed) and half an hour waiting for the train.

For a business traveler like me, this train trip would be much better than a plane trip because I would be more productive.  I can only get about an hours' worth of work in if I fly.  The rest of my time is wasted driving to the airport, going through security, etc.  On the other hand, I would have three hours of productive time on a train.  Plus I'd be more comfortable.  I would gladly trade a little longer trip for that productivity and comfort.  And I would be willing to pay much more for that train trip.  If I got to keep most of that surplus, and there were enough other travelers like me, high-speed rail might make sense.

But this new proposal would eliminate much of high-speed rail's advantages.  By requiring airport-to-airport travel, it would add at least 40-50 minutes to the trip -- more, if it was necessary to go through airport-security.  The plane's time advantage over the train would grow from, say, one hour to almost two hours.  And that time added to the train trip would be unproductive.  Rail travel would be much less attractive for business travelers.  This no doubt explains Southwest's "neutrality" on the proposal.

Claiming the rail would be a "feeder" for interstate flights is silly.  We already have feeders for interstate plane travel.  They have wings and flight attendants and park at the terminal gates.  High-speed rail would have no advantage over planes as a feeder.

It's a good time to renew the debate over high-speed rail in Texas.  But the debate needs to begin with a proposal that makes sense.

The Postal Service and those red envelopes

The U.S. Postal Service might cut out one day of service a week because it's going bankrupt.

I really wouldn't care . . . except for Netflix.  It's not like I watch a movie every night, but when I do want to watch the next flick in my queue I don't want to wait four days for it.  

I imagine the Netflix execs are sweating this one.  Sure, Netflix could cut its labor costs by a sixth if the post office were to cut out a day of service.  But I'd be less satisfied with the Netflix set up and would probably go back to Simpsons and Family Guy reruns.  And if I'd hate losing that sixth day of service, I'm sure most other customers would too.  I bet Netflix would gladly keep paying for six days of labor rather than have the Postal Service screw with its successful business model. 

Freeriding

In 1998, Cedar Park voted to secede from Capital Metro's service area.  Despite having shucked their contributions to Cap Metro, many Cedar Park residents exploit Cap Metro by driving up to Leander to catch Cap Metro's express buses.

Cedar Park is now trying to devise a system to help its residents freeride:

Now city officials are looking for ways to pay for a proposed transit system that would shuttle residents around town during peak hours and stop at one of Capital Metro's park-and-ride lots, where they could get transit service to Austin.

Cedar Park officials, of course, have no plans to join Cap Metro's service area -- it's much cheaper simply to drive its residents to Leander.  And, not content merely to freeload, Cedar Park wants to grab some of Cap Metro's federal dollars, on the theory that it's entitled to its "fair share."

Cap Metro has tolerated this for far too long.  I suppose it believes that sorting Leander residents from Cedar Park residents is just too hard.  But it really shouldn't be.  

I suggest setting shuttle bus fares at full cost, including capital and overhead expenses.  That fare might be, say, $6 one-way.  Then offer deeply discounted fare cards to anyone who can establish residence in Cap Metro's service area.  Providing a Leander mailing address when ordering a card online would be good enough.  Cedar Park residents could then pay their own way and Cap Metro taxpayers could get the discounted rides they (mostly) pay for.