Monday, September 24, 2007

Digging into the newest congestion numbers

So the Texas Transportation Institute at TAMU came out with their 2007 Urban Mobility Report last week. In case you missed the story in the Chronicle, it's steadily getting worse in Houston, now up to 56 hours of average delay per traveler per year, which is 7th worst in the nation. On the plus side, we're only 2 hours/year over the average for very large cities (54), and only 12 hours/year over the national average of 44 hours for the 85 cities studied. That's less than 3 minutes of penalty per workday for commuting in Houston vs. the average major US city.

Some other interesting points I found in the report:
  • Multi-billion dollar transit systems don't seem to help alleviate congestion too much: DC and SF are worse than Houston, and NYC, Chicago, and Boston are each only slightly better, with 46 hours of annual delay. NYC and Chicago actually have a worse travel-time index than Houston (ratio of trip times at rush hour vs. off-peak), with 1.39 and 1.47 vs. our 1.36.
  • In the main report, Exhibit B-36 pg. 130 (B-52) has a graph showing not just Houston's change in travel times through the day, but how much margin time you have to put in your travel planning to be on-time for 19 out of 20 trips. It's very substantial, and a lot of that time goes to waste, as people arrive early at appointments. This is a strong argument for congestion-priced toll lanes, which give people the option of a guaranteed fast trip when they need it. People who need to be somewhere at a specific time can pay for it, and save the time they would otherwise waste leaving extra early "just in case." People on a more flexible schedule can stay in the free lanes.
  • "Can more road space reduce congestion growth?" Their answer is essentially 'yes':
"The analysis shows that changes in roadway supply have an effect on the change in delay. Additional roadways reduce the rate of increase in the amount of time it takes travelers to make congested period trips. In general, as the lane-mile “deficit” gets smaller, meaning that urban areas come closer to matching capacity growth and travel growth, the travel time increase is smaller."

That same appendix also has a graph on pg.3 clearly showing that cities that made a better effort to keep capacity in line with growth (such as Houston) experienced far slower growth in congestion than those cities that let it get away from them. In fact, when ranked by how much worse our congestion is vs. 1982 (when the study started), we fall dramatically to 27th, which reflects our strong capacity increases over that period.
  • Our I-10 project gets featured on pg. 23:
"Constructing transportation projects quickly and with as little extra delay as possible requires a mix of strategies, just as the regional approach to congestion relief. The Katy Freeway (I-10 West in Houston) expansion project includes additional mainlanes and high-occupancy toll lanes, in addition to reconstructed pavement, noise walls and landscaping. The regional toll authority purchased the right to operate the toll lanes using funds generated over almost two decades of successful toll operation in other corridors. The accelerated cash flow enabled the Texas DOT to decrease the construction time from 12 years to six years. The increased cost of the 24-hour construction schedule was partially offset by savings in construction cost inflation that would have occurred. The estimated $2.8 billion in benefits that resulted from six years of improvements in delay, lower fuel consumption and improved business environment more than offset the estimated $300 million in extra costs."
I'd love to see that kind of math applied to more projects in Houston.

One of the points I keep hearing over and over laments that "if we build more road capacity, it'll just fill up," implying we shouldn't bother. I believe the Chronicle editorial board even made this point, although I can't find it online to confirm. This argument really bugs me. Do you ever hear people say, "Ugh, why build more airport runways - the airlines will just fill 'em up with flights," or, "You know, if we expand the port, it'll just fill up with more ships and trade."? No, you never hear that, because people realize those are good things to have happen. Yet, somehow, expanding road capacity to enable more economic activity and better access to jobs and affordable housing for more people is a bad thing? I don't get it.

Labels: , , , ,


At 10:06 PM, September 24, 2007, Blogger John said...

Well, the actual argument is a bit different than that (which is why, in the form you present it, it indeed does not make sense).

Roads tend to generate development beyond their capacity to carry resulting traffic, thus the road becomes a subdisy to development which is going to create congestion and eventually require more transportation investment.

All of this of course creates other costs which are borne by the community at large - loss of land for other uses, air pollution, and so on - all of which are in effect development costs which the people building and using the new development don't pay (the rest of us do).

Which is the problem with the personal auto transportation paradigm - for all the great things about it (and yes, they are great, and I am one of the users of it) when it expands too much, it causes a whole bunch of other problems.

One way to address this is building roads with much more limited access - exits every five or ten miles, for example, with development concentrated there - so that they serve as efficient feeders between activity center, but then the people who have the road running through their communities but no use of if have a valid complaint about it.

The problem is really that costs are not born by users. We subsidize development at the edge of urban areas by running highways through established neighborhoods.

At 10:25 PM, September 24, 2007, Blogger Tory Gattis said...

Actually, the subsidy seems to be quite small. See:

That analysis also completely ignores the extra property taxes generated by the development, which can further mitigate any potential subsidy.

Clearly, value is being created, because people do rapidly by into these developments and use these roads. The question is how that value compares to the costs, and who bears those costs relative to who gets the value. I'd bet the net value is substantially positive, but I'll admit its a very complex and subjective calculation.

At 11:24 PM, September 24, 2007, Blogger Michael said...


Nobody says "don't build transit", which you equate to building road roads. People may say, "invest in transit wisely". Your comparison to air travel is a good case in point - there are many forms of air travel for which there are several different solutions that we invest in as taxpayers. This involves everything from the space program to medical helicopters to the air force, spy satellites, etc. Adding extra runways to IAH and Hobby does not solve all air transportation needs, much as expanding I-10 does not solve all our congestion problems. Actually, from what I've read of the study, the authors repeatedly point out that this problem is complex, and, like air travel, requires a complex strategy - there is no "one stop solution" or quick fix.

The argument that expanding highways yields merely more congested highways seems to be a pretty good argument for looking at commuter rail etc. to me. Looking at NYC / Chicago's better delay numbers, in spite of 3-5x the metro population (which would suggest that they would have much higher transit times, all else being equal), is further confirmation that a transit strategy involves more than roadways.


At 11:42 PM, September 24, 2007, Blogger Michael said...

Furthermore, if the issue is "intercity travel" when you are referring to air travel, again, the answer is not always to "build more runways" at IAH. Cases in point:

- Amtrak has 1.2 million riders now annually in the SF - Sacramento line (when they could have just "built another airport runway" I suppose?)
- Suburban airports / private jets are often a better answer for some customers - witness their proliferation
- The air traffic system is congested due to Air Traffic Control outdated equipment and delays at hubs- building more runways in Houston does not help a flight from Newark take off any sooner
- Airbus / Boeing are racing to add more capacity per flight with new airliners

So, I like your analogy. It is a good example of how closed your thinking is - clearly "build another runway" is about as good an answer as "build another roadway" - to relieve congestion issues. That is, it is a naive solution that only helps some of the time, in some cases.

At 9:05 AM, September 25, 2007, Blogger Tory Gattis said...

I didn't say capacity additions are the only solution. But they are a good one. And I'm actually familiar with the air travel issues, and most experts say that there are various measures that will help (inc. redesigning how we do air traffic control), but at the end of the day runways are desperately needed in popular destinations like NYC.

Commuter rail might make sense to look at as a solution when the HOV lanes are packed with full express buses and approaching capacity, but since they're no where close to that level (nor on any growth trend towards that level), why invest billions in the capacity addition of commuter rail when the demand is clearly not there? (while the demand *is* clearly there for freeway lanes)

At 9:23 AM, September 25, 2007, Anonymous Anonymous said...

Congestion pricing huh? By that logic, I should be paying roughly .20 cents to use Hardy Toll Road southbound at 5pm right?

When Houston agrees to that, they can double their rates on Westpark.

At 9:27 AM, September 25, 2007, Blogger Tory Gattis said...

Actually, I think a discount in the contraflow direction makes total sense. Even better would be to have more lanes be one-way reversible in the rush direction.

At 10:29 AM, September 25, 2007, Blogger Brian Shelley said...


I think that John is vaguely correct. “Free”ways do subsidize suburban development even if it is, as you admit, small. I observe that people have an infinite marginal utility for almost any item that’s free. I’m pretty sure you could hand out free samples of dog vomit and people would push and shove to get it. There’s a massive marginal loss of utility to charging for something that used to be free, thus moral outrage.

When a freeway expands it changes behaviors that cause congestion to persist. Many people wake up early to beat the worst of the traffic. If you expand capacity they will start getting up later in a tradeoff between congestion and sleep. The converse is true for those who get to and leave work late. Less congestion attracts former non-users adding again to congestion. Because of this trade-off between preferences the users of freeways don’t see much improvement personally (regardless of data) and believe the more capacity <> less congestion equation.

A tollway does not have the same behavior dynamics. If the profits are maximized then subsidization falls near zero. Congestion can be limited by congestion pricing. The pricing encourages carpooling and bus/transit use. The problem also exists for local roads, but tolling is too much of a hassle, so other taxes are a good enough proxy.

There is no problem with the personal auto transportation paradigm, only the free road paradigm.

At 1:40 PM, September 25, 2007, Blogger Michael said...


I think that the freeways being packed, not merely HOV lanes being packed, is also a good measure of whether you have a candidate for commuter rail. And the HOVs are generally pretty crowded, FWIW.

The argument for "look beyond roads" is that the current cycle is Expand -> Some Relief -> Congestion -> Expand -> and the cycle repeats. The problem is, as this cycle repeats, right of way costs and required space increases. In some cases, right of way costs alone may make a project pretty much unfeasible. And land is a scarce resource - they aren't making any more if it, except in volcanic regions, etc.

For example, while expanding I-10 is great and should solve many problems over the next several years, I don't think I-10 will be expanded again between Highway 6 and 610 in our lifetimes, and if we do expand it, it will come only at an astronomical cost. So, the question is, 10 years from now when we are once again facing increasing congestion on I-10, what is our strategy? Expand the highway again?

If you look at the example of airports, many cities would love to just "add a runway" - but this is no longer a practical solution because they do not have space, or would have to make right-of-way acquisitions and buy up very pricey developed land. Fortunately in Houston we don't really have that problem with our airports or our roadways... yet.


At 2:18 PM, September 25, 2007, Anonymous Anonymous said...

[quote]Multi-billion dollar transit systems don't seem to help alleviate congestion too much: DC and SF are worse than Houston, and NYC, Chicago, and Boston are each only slightly better, with 46 hours of annual delay. NYC and Chicago actually have a worse travel-time index than Houston (ratio of trip times at rush hour vs. off-peak), with 1.39 and 1.47 vs. our 1.36.

Of course they don't alleviate congestion too much. That's not the point. The point is that they give people the option of avoiding the congestion. So instead of having to get onto a clogged freeway system, you can go to all the major destinations in your city by train.

NYC and Chicago are only slightly better than us delaywise? Well then they must be doing something right, because they are handling populations of 20+ million and 9 million (respectively) in a land area no bigger than we use to handle 5 million. Of course, if they had gone the car-only route and just added freeway capacity for all those people, you would have 30 lane freeways, and probably half of that land area would be concrete.

At 2:44 PM, September 25, 2007, Blogger Tory Gattis said...

Brian: Agreed.

Michael: I agree there are practical limits on expanding the freeways. When those are reached, it's up to congestion pricing on at least some of the lanes to manage demand and encourage transit and car/vanpooling. Even better would be to acknowledge that there is little demand in the contraflow direction, and convert several lanes in the middle to one-way reversible congestion-priced toll lanes. Along many of Houston's spoke freeways, we could get to 8+ lanes of capacity available in the demand direction, several of it priced to guarantee high speeds, including for transit.

> Transit: The point is that they give people the option of avoiding the congestion.

David Crossley made the same point to me in email, and I agree. But "congestion reduction" is held up as a reason for many transit projects, and it's simply not true. I also argue that express lanes+buses provide that same option at far lower cost than commuter rail.

As far as additional runways: they are an easily affordable option relative to the many billions in airport revenues from passenger fees. They are disruptive to the local neighborhood, which is why they are rare - but clearly the regional benefits outweigh the local costs. The problem is political fear at the vocal neighborhood, as well as ambivalent airlines, because they actually realize that limited capacity improves their ability to raise fares (constrained supply with rising demand).

At 5:42 PM, September 25, 2007, Anonymous Anonymous said...

Why would anyone think that adding rail lines would be less disruptive than adding highway lanes? An effective rail system covering a long route is going to have to incorporate a four track local/express/express/local alignment, will be loud and will require parking, signaling and other heavy infrastructure of its own, most of which is alternatively loud and idle. All the legal arrangements put in place over the years to stall and prevent highway construction is available to stop rail also, and you can find vociferous opposition on pure NIMBY grounds to rail corridor work of the same intensity as that marshalled against highway work.


The principal reason that Amtrak is so slow on the Northeast corridor between NYC and Boston is because it has to run on an ancient alignment with severe curves and grade crossings over a narrow RoW. Acela could do 150 mph if it could run in a straight line without grade crossings over its own dedicated tracks. The NIMBYs in New England don't want that, and so you have a slow service, presumably in perpetuity.

Also: One unfortunate outgrowth of NIMBY and related opposition to highway work, is that designers seem stuck in having to manage or add capacity exclusively along existing alignments rather than design more narrow, and more attractive roads at decent intervals. If you had a few highways with RoWs of, say 6 lanes (3 per direction), you might not need an 18 lane behemoth running down the gut of the city.

At 7:44 PM, September 25, 2007, Blogger Tory Gattis said...

I absolutely agree with that last point. Most cities, including Houston, didn't build the freeway grid close enough together, so we end up with mega-freeways instead of more manageable sizes. Something like a Memorial or Allen Parkway on a 1-2 mile grid would have been nice. But at this point we're pretty stuck with expanding what we have. I hope at some point roads like Memorial or Allen Parkway might go in some of rail corridors, or corridors of the heavy power transmission lines. But even if they're low key like that, the NIMBY opposition will be large.

At 9:01 AM, September 26, 2007, Blogger ian said...

"NYC and Chicago actually have a worse travel-time index than Houston (ratio of trip times at rush hour vs. off-peak), with 1.39 and 1.47 vs. our 1.36."

Do these indices encompass only those trips made by automobile, or do they also include transit trips? I'd be very surprised if it took a subway rider more time to get from point A to point B during the peak hour, since that's when they have the most trains running. And with how tightly I've seen some riders pack onto subways, I doubt many people have to wait for more than one train.

In fact, I'm going to assume for the moment that these travel-time indices DON'T account for transit trips. If that's the case, then this point is totally moot since Chicago and particularly NY have made concerted efforts to focus more on transit and less on highways. If transit users aren't included in the index, then the numbers will be very skewed.


At 11:53 AM, October 04, 2007, Blogger M1EK said...

Don't trust the congestion figures for cities with high transit usage - they basically assume that we should only care about motorists, even when half of commuters take transit (as in New York's case; lesser degree of impact on DC and SF where substantial people aren't affected at all by the motorist congestion yet the area gets marked as "bad traffuc").

At 11:56 AM, October 04, 2007, Blogger M1EK said...

"I also argue that express lanes+buses provide that same option at far lower cost than commuter rail."

It provides that option only if passengers are equally willing to take the bus trip, which they are provably NOT, even with good bus and poor rail service. Those buses eventually merge with regular traffic, at which point they become slow, poky, city buses, which choice commuters avoid like the plague.

At 12:06 PM, October 04, 2007, Blogger Tory Gattis said...

The congestion figures are accurate even for transit-heavy cities. It just shows that even massive transit systems don't relieve congestion much, and are inconvenient enough that most people still choose to drive even with the congestion.

The street grids within the job centers are not generally the problem - it's the freeways. And commuter rail will still stop far from most people's final destinations (note the intermodal terminal plan north of downtown), so they'll still have a transfer+slog of some sort.


Post a Comment

<< Home