Inducing Demand

This Washington Post article on severe traffic problems in Northern Virginia is perhaps a good opportunity to try to explain the concept of "induced demand." I feel like transit advocates sometimes talk about this like it's a mysterious phenomenon where you build the roads and, mystically, the road calls into existence its own traffic. Realistically, it is, of course, true that if you widened the major NOVA highways the traffic would start flowing faster on them.

But you need to consider what happens next. Right now, many Virginians commute to work by taking Metro or the VRE commuter rail or one of several buses. If there were less traffic, fewer of those people would be inclined to do so. Similarly, right now the very long commute is one reason people might not want to move to certain parts of Prince William or Loudon Counties. But if the commutes got shorter due to reduced traffic, them more people would move. Soon enough the roads are congested again. This isn't necessarily a bad thing, it just illustrates the point that the purposes of building new roads is to expand the area under development not to relieve traffic congestion. Small dirt roads rarely feature traffic jams, not because they have high capacity but because nobody wants to drive on them. You replace a small dirt road with a better paved one not to improve the flow of traffic, but to encourage people to locate stuff -- homes, businesses, etc. -- along the route of the road. The goal, in short, is to induce demand for driving on the road.

At the end of the day, an uncongested road is an excellent thing and driving on one is a fast and convenient way to get places. And when you have a valuable commodity that's not priced, over the long run people wind up over-consuming it and creating shortages. If the government set up a "french fries trust fund" to cook that were then given away for free, they'd soon enough run out of french fries. It's the same with the effort to build uncongested highways. Over the long run, it only really works for absurd pork barrel schemes that involve building giant roads in places nobody wants to go. If you want an uncongested road in a desirable location, you need to put a price on using the road during peak times and then you need to take the funds raised by the congestion charge to finance alternatives.

Photo by Flickr user ethandb used under a Creative Commons license

Comments

The short version is that people have a level of tolerance for congestion. When congestion is below that tolerance level, they will be more inclined to drive on those roadways until congestion passes the tolerance threshold, at which time they will moderate their driving habits or move to a new place to keep the time spent in congestion in line with what they can put up with.

You could call it "The Iron Law of Traffic."

Posted by Tyro | May 5, 2008 1:33 PM

Sure, put your lot in with economists!

Posted by Chris | May 5, 2008 1:35 PM

If you want an uncongested road in a desirable location, you need to put a price on using the road during peak times and then you need to take the funds raised by the congestion charge to finance alternatives.

I agree with this in theory but the problem is that people in Prince William and Loudon aren't commuting to the District -- they're mostly commuting to other parts of Northern Virginia. The drivers that make the NOVA part of the Beltway an ordeal at rush-hour aren't headed for your neck of the woods. Thus, the alternatives you are describing are difficult to envision and prohibitively expensive.

For someone who lives in, say, Woodbridge and works in Falls Church, driving is the only realistic way to get to work. Even if there was a bus between home and work, it certainly wouldn't provide the flexibility that people with kids need.

I can imagine mass transit links between Loudon and Tyson's Corner but it's going to have to be a rail one -- why sit on a bus in Route 7 traffic when you can be in your own car in Route 7 traffic? What I can't imagine how expensive such a link would be and who would be willing to pay for it.

Posted by Roberto | May 5, 2008 1:38 PM

Why not take the money raised and fund higher education, then perhaps we could produce some advisors who would advocate for something maybe a bit less incompetent.

Posted by Aaron | May 5, 2008 1:40 PM

I'm not sure if this is as simple of a relationship as you make it out to be. I live in Austin, for example and I know certain roads which are really bad to travel on during the wrong hours. At the same time, there are plenty of other more reasonable routes to take and it's possible to structure your work and living situation such that you can avoid traffic. At the same time, there are particular areas of congestion that create bottlenecks that make what would otherwise be reasonably short commutes much longer. It may be true that if you build a road from the city to the suburb, that road will always be at maximum capacity. But there are also smart ways to figure out where to add road capacity so that it doesn't take an hour to get over the bridge into the city, which is a problem whether you live 30 miles away or just across the river.

Posted by mpowell | May 5, 2008 1:43 PM

Clip this post out and put it alongside your mirror. When you wonder why people think you are a young, unencumbered, overpaid, elitist pinhead, just walk next to the mirror and pop a few more zits.

Posted by jerry | May 5, 2008 1:44 PM

One might argue that the transit advocates of which you speak do not, in fact, hold a strange mystical view in which roads magically produce traffic to fill them, but rather that they understand the phenomenon you sketch here but use a shorthand description of the phenomenon that omits certain events in the causal chain.

Posted by djw | May 5, 2008 1:46 PM

Put it next to your toilet and you can rub one out while reading your own libertarian porn.

Posted by jerry | May 5, 2008 1:46 PM

The other side of the argument is that if you really want people to use mass transit, you're going to have to increase traffic congestion and reduce road capacity. And... how likely is that?

Posted by MattF | May 5, 2008 1:53 PM

If the government set up a "french fries trust fund" to cook that were then given away for free, they'd soon enough run out of french fries.

Analogies like this are supposed to clarify things, not make them incomprehensible.

Posted by washerdreyer | May 5, 2008 1:53 PM

A good example in my Tacoma area: There's a great deal of suburban development across the Tacoma Narrows, which is a narrow part of the Puget Sound. Traffic over the Tacoma Narrows bridge had become intolerable, so a new bridge was recently built. I was cynical about this for exactly the reasons you mention; I figured the existence of the new bridge would simply lead to even more development, and in 5-10 years we'd be back where we started. However, the WA government is charging tolls for the new bridge-an example of forcing people to bear some of the cost-so maybe that will help. I sure hope so.

Posted by beckya57 | May 5, 2008 1:56 PM

But if the commutes got shorter due to reduced traffic, them more people would move. Soon enough the roads are congested again. This isn't necessarily a bad thing, it just illustrates the point that the purposes of building new roads is to expand the area under development not to relieve traffic congestion.

Your "transit good, roads bad" arguments are getting more and more confused. There may obviously be more than one purpose for building new roads, including both relieving congestion and stimulating development. And just because building a road has the effect of stimulating development obviously does not mean that was its purpose or its primary purpose. And there are obviously constraints on the growth of communities other than commuting times and the quality of their road connections. The idea that whenever you build a new road to relieve congestion it will induce new demand that will produce the original level of congestion simply isn't true as a factual matter. Numerous cities and metro areas have seen reduced commuting times as a result of new and expanded roads.

Posted by Mixner | May 5, 2008 1:57 PM

A good example in my Tacoma area: There's a great deal of suburban development across the Tacoma Narrows, which is a narrow part of the Puget Sound. Traffic over the Tacoma Narrows bridge had become intolerable, so a new bridge was recently built. I was cynical about this for exactly the reasons you mention; I figured the existence of the new bridge would simply lead to even more development, and in 5-10 years we'd be back where we started. However, the WA government is charging tolls for the new bridge-an example of forcing people to bear some of the cost-so maybe that will help. I sure hope so.

Posted by beckya57 | May 5, 2008 1:58 PM

I'm actually one of those people who lives in Loudoun and commutes into DC. I usually take the Loudoun Commuter Bus. It's an excellent service that makes a couple stops in Leesburg and points north, then goes along the Greenway/Toll Road to 66, then into the District through Rosslyn. It takes about an hour and ten minutes. Regardless of how clogged or fast-moving the toll road gets, I'd still take the bus - because it costs less than the gas plus the tolls would be.

(Incidentally, when I was driving in to work, I discovered that the Greenway's HOV lane cost less than taking 7 in all the way, because of all the stop lights on 7).

Posted by Tel | May 5, 2008 2:33 PM

1) More efficient land use. One way is to get rid of zoning laws entirely so that NIMBY cities can no longer set ridiculous, sprawl-creating, density-reducing minimum lot sizes. Another method is urban growth boundaries; Canadian cities as well as Portland, OR, have more affordable housing and shorter commutes than many comparable US cities. If they need more land to develop, they can always expand the growth boundary; cities without growth boundaries have already used up all their accessible land, and housing grows more expensive. Maybe combine the two -- simply set a growth boundary and no other zoning law?

2) Let the market actually bear the cost of development. Don't burden taxpayers with the cost of new roads and new sewers, let the developer pay for them. That will get them thinking more efficiently very quickly as well as lowering real estate taxes.

3) Change the law to make infill development or brownfield reclamation easier.

4) Pay more attention to revitalizing close-in neighborhoods. Fix the schools. Stop the teachers' unions from protecting bad employees and make them accountable. Provide low-interest loans for people willing to take a chance on fixing up an old house in a sketchy neighborhood. Get serious about crime prevention instead of depending solely on deterrence through prison time. After all, nothing drives sprawl more than the urge to escape crime and especially bad schools. Schools are the biggest single factor in driving where families choose to live; there is almost no limit to what people with children will put up with in terms of commuting and house prices if a good school is set out for them like some kind of big juicy carrot.

Posted by DBX | May 5, 2008 3:28 PM

Sorry folks, I missed the first paragraph off the post. That should read:

If you want more viable transit and less sprawl, you need to have more compact and dense cities. European cities, British cities, even Canadian cities are more compact and have higher transit use than American ones do. Remarkably, Canadians have double the per capita transit use and on average live closer to one another than Americans do. But how to achieve this outcome? Some suggestions:

Posted by DBX | May 5, 2008 3:34 PM

Oh dear, not having a good editing day here. "If you want less congestion, more viable transit and less sprawl . . . . ."

That should do it for the first line.

Posted by DBX | May 5, 2008 3:37 PM

There may obviously be more than one purpose for building new roads, including both relieving congestion and stimulating development. And just because building a road has the effect of stimulating development obviously does not mean that was its purpose or its primary purpose.

Relieving congestion can't be a reason for building more roads any more than wanting to lose weight could be a reason for eating Big Macs every day. The chosen action is counterproductive toward the stated goal.

The idea that whenever you build a new road to relieve congestion it will induce new demand that will produce the original level of congestion simply isn't true as a factual matter.

Really? You've got data that supports this? You might want to share it with the entire transportation planning field, because they disagree with you.

Posted by Phaedrus | May 5, 2008 3:47 PM

A one-stop ultra-easy-to-administer way of discouraging over-reliance on cars is to simply raise gasoline taxes. At $8/gallon those way-out-in-the-sticks developments won't look so attractive, even without congestion. At $10/gallon people might even think harder about abandoning central cities.

Awhile back I read a book called Edge City by Joel Garreau, which mentioned that we have not built a downtown anywhere on earth in the past 70 years or so, instead doing all development in the low-density suburban mode. Once petroleum becomes expensive enough we might rethink this.

Posted by jimBOB | May 5, 2008 5:20 PM

Numerous cities and metro areas have seen reduced commuting times as a result of new and expanded roads.

As Matt alluded, there are many, many places where this is true in the short term. It's even possible the effect could last longer than 5 or 10 years if, say, the regional population is holding flat or declining, or has a powerful senior senator and is enjoying some kind of sustained federal freeway largesse.

In the long term, though, it will catch up. Curing congestion with more roads is like curing obesity with larger pants.

I'm not sure I can think of any metropolitan area where the supply of suburban land is very restricted once you get out 70, 80, 90 miles from the outer ring of heavy urbanization, so I don't really know what you mean by "other constraints".

Obviously, other things that might restrict development are rising gas prices, land use and environmental regulation, and even changing consumer preferences. But that can't be what you mean, because you seem to react quite badly elsewhere to any suggestion that those things could or should result in any slowdown to sprawling auto-oriented suburban expansion...

Posted by jack lecou | May 5, 2008 5:28 PM

Of course it doesn't help that the Defense Dept is moving 20,000 jobs from mass transit served offices to Fort Belvoir which is at the end of the 2 lane "choked" secondary road that is highlighted in the article - the road that the folks down in Richmond have now decided not to enlarge.

Regardless of any arguments about "Inducing Demand", nothing make sense when decisions are dictated top-down by non-local bureaucrats.

It's not like widening Rolling Road is going to "create" more traffic - that traffic is going to be there whether the road gets widened or not - due to some extremely stupid governmental policies.

Posted by Ethel-to-Tilly | May 5, 2008 5:44 PM

Phaedrus,

Relieving congestion can't be a reason for building more roads any more than wanting to lose weight could be a reason for eating Big Macs every day. The chosen action is counterproductive toward the stated goal.

This claim is utterly nonsensical.

Really? You've got data that supports this? You might want to share it with the entire transportation planning field, because they disagree with you.

Yes, really. Neither of your sources says what you seem to think it says. The Census Bureau reports that the average commute time in the United States declined slightly between 2000 and 2005. In some places, commuting times increased, and in others they decreased. One area where commuting times have declined is California. The median commute time in California declined by 9% between 1990 and 2004.

Posted by Mixner | May 5, 2008 6:32 PM

Also, the idea that higher densities and more transit will produce shorter commutes is a fantasy. As reported here, the average commute time in Europe is substantially longer than in the U.S. European have the worse of both worlds: smaller homes and longer commuting times.

Even in Britain, which has terrible road congestion and far more public transit than the U.S., the number one reason drivers gave for driving to work rather than using other options is that driving is quicker. And when asked what they would do if their driving time doubled, only 7% said they would switch to public transit.

Posted by Mixner | May 5, 2008 6:50 PM

DBX,

Another method is urban growth boundaries; Canadian cities as well as Portland, OR, have more affordable housing and shorter commutes than many comparable US cities.

Debunking Portland: The City that Doesn't Work

Posted by Mixner | May 5, 2008 7:02 PM

And silly glibertarian boy Mixner heads off to Cato, where they surely haven't got an axe to grind. That's really not going to impress Megan: be more creative, silly boy.

Posted by pseudonymous in nc | May 5, 2008 8:51 PM

A one-stop ultra-easy-to-administer way of discouraging over-reliance on cars is to simply raise gasoline taxes. At $8/gallon those way-out-in-the-sticks developments won't look so attractive, even without congestion. At $10/gallon people might even think harder about abandoning central cities.

Racial minorities not being so prone to committing crime would also go a long way to getting more people to ove back to the cities.

This claim is utterly nonsensical. -
-Mixner to Phaedrus

I think, Mixner, that Phaedrus was being sarcastic.

Posted by Glaivester | May 5, 2008 9:14 PM

Im sorry, but Matt's dogmatic understanding of transit lacks any knowledge of history. It's not as if people are all of a sudden choosing to live really far out. Maybe if you took urban studies classes rather than philosophy (grounded in theory) you might realize that suburbanization was incentivized as was (and is) housing development. We can't just create a world that does not exist. Sprawl is here and you can't use libertarian ideological convictions to erase the past. Matt posits a United States absent any history and sees congestion pricing as a panacea to all woes transportation.

You grew up in New York and live in Washington D.C., two cities with strong, mature (somewhat in DC) transit systems. In many other places land use is not so dense and transit not so useful. Taxing people that are already stretched thin is akin to, absent a better phrase, "blaming the victims."

Stop using a one size fits all formula for every municipality. It is sort of reminiscent of exporting free market, secular democracy abroad. It is probably better to contextualize. Well, I guess Iraq is pretty good.

Posted by jeff | May 5, 2008 9:24 PM

Jeff,

There is no one right answer, to be sure, but there sure are lots of wrong answers. Opposing the construction of mass transit systems is one of them.

Here in Los Angeles we've got our history with and without mass transit. Not just the Pacific Electric which was mothballed to make way for automobiles, but the 1985 Congressional legislation that banned the construction of subways on the westside, legislation that was repealed in 2007 amid an increasing gridlock visited upon those well-heeled neighborhoods by their short-sighted nimby-ocity.

Today it can take 40-60 minutes to traverse the single mile from Bundy and Olympic to Sepulveda and Olympic (one block north of the famous Pico and Sepulveda memorialized in song).

What's great is that the same congressman who got kudos for saving the westside from the danger of subways got kudos for repealing his own bill, and he reprepsents the San Fernando Valley, not the people he screwed over with his landmark legislation. Here's to you, Henry Waxman, thanks for nuthin.'

Posted by west coast | May 5, 2008 10:51 PM

"One area where commuting times have declined is California. The median commute time in California declined by 9% between 1990 and 2004."

And how much did the California freeway system increase in those years vs. how much the various forms of public transit increased over that period?

Posted by Adam Villani | May 6, 2008 1:34 AM

And how much did the California freeway system increase in those years vs. how much the various forms of public transit increased over that period?

Did you read the link? The drop in commute times wasn't due to public transit.

Posted by Mixner | May 6, 2008 2:36 AM

"Stop the teachers' unions from protecting bad employees and make them accountable."

Not to drag the conversation off-topic, but the meme of 'teacher's unions protecting bad employees, etc.' is somewhat misguided. - One almost gets the impression of teachers' unions running around specifically trying to protect bad employees for some bizarre & counterproductive Now, I'm not saying that it's not an imperfect situation, no doubt with genuine problems & abuses on occasion - but this general accusation kinda misses the point, possibly because the very idea of unions has become so unfamiliar and alien to many folks. (Also because teachers are supposed to work out of the blessed goodness of their soft little caring hearts and when they don't - what with it being a largely female quasi-family job -it can set off anxieties related to caretaking issues, etc.)

The job of a union is to provide some measure of protection (and ideally benefit) for its members. There are cases where unions stand up for genuinely bad employees. That's because unions are standing up for procedures and safeguards intended to protect teachers as a whole, and can't be suspended because one feels like it - even for a good reason - without seriously weakening the whole thing. (Again, this is an alien concept for many, because their workplaces have been stripped of (or preventing from ever achieving) these kind of basic protections. Really basic starting point: do you believe your boss is your friend? Do you think their judgement is perfect and impartial? Even if that is the case in your specific situation, do you think it applies generally?). Often union reps are pretty clear on who the 'bad teachers' are, and generally want them out as much as anyone. But one has to follow procedures, and there are cases where honestly, administration just doesn't want to make the effort. Again, I'm not saying the system's perfect - not at all - and surely there are specific things that go too far, and could be usefully reformed. But the point of teachers' unions is to avoid things like this, where a substitute teacher may be fired because he was accused of wizardry (for performing a quick magic trick - involving a disappearing toothpick - in class). Whether or not there's more to the story, since he's an at-will employee, it's entirely possible for him to get fired for wizardry. Unions and other such protections make it harder for folks to get fired for things as ' some parents are complaining that the science teacher is indoctrinating the children with Godless evilutionism', or 'because the principal is playing petty politics and teacher B isn't part of their power base' or whatever; it's an unfortunate side effect that it also means that firing teacher C who should really have quit 5 years ago and has basically given up will require jumping through some hoops (again, I'm sure in some cases unnecessarily high ones.)

Posted by Dan S. | May 6, 2008 7:28 AM

Mixner,

The CA study says median commute times dropped because jobs moved to the suburbs.

So a 30 mile commute becomes an 8 mile commute, and the median travel time drops.

But if the 8 mile commute now takes twice as long as it used to take, then the average time could increase even as the median decreased.

It's called "lying with statistics," a time-honored way for people to see only what they wish to see.

Posted by west coast | May 6, 2008 4:42 PM

west coast,

The CA study says median commute times dropped because jobs moved to the suburbs.

No, the study doesn't say there was just one reason, but the author suggests that the primary reason was an increase in the share of suburb-to-suburb commutes and commutes starting in suburbs and ending in cities. People and jobs both moved to the suburbs, and roads connecting cities to suburbs and suburbs to other suburbs facilitated this change. Public transit had only a negligible impact on the reduction in commute times. The median commute time fell by 9%, but transit use increased by only 0.2%.

So a 30 mile commute becomes an 8 mile commute, and the median travel time drops.

I don't see any claims in the report about changes in the median (or average) commuting distance.

But if the 8 mile commute now takes twice as long as it used to take, then the average time could increase even as the median decreased.

The statewide average did increase (did you read the article you're commenting on?), in part because of an increase in very long commutes. But those are the minority. The statewide median is more representative of the commute for most workers, and the median declined, as I said. And the average also declined in the five most populous Bay Area counties.

It's called "lying with statistics," a time-honored way for people to see only what they wish to see.

You're the one who is lying. You're attributing claims to the study it does not make. And you also seem to be confused about the meaning of statistical measures, confusing averages and medians.

Posted by Mixner | May 6, 2008 5:05 PM

Oh, geez.

Ok, you caught me. By inadvertently agreeing with your article and claiming that average commute times are up while median commute times are down, I am unmasked as a LIAR!

You win. The world's a better place now than it has ever been, and mass transit is no solution to our transportation needs, and etc.

Now I'll climb into my car to drive the 12 miles to my home in one hour, a trip that once took only 30 minutes, and thank my lucky stars that Mixner has pointed out the error of my thinking when it comes to time spent listening to NPR.

L8TR

Posted by west coast | May 6, 2008 9:09 PM

As usual, Mixner throws up a lot of gravel, while managing to miss the point almost entirely.

On the two-dimensional surface of the earth, raw statistics on commute times have absolutely no bearing on the phenomenon of induced demand. At minimum, you need to control for new road construction (at minimum - changes in residence and work location also come into play, so it's pretty much completely useless.)

Nor do they show anything useful about the utility of rail or other public transit.

Commute times are based on a whole complex of public decisions regarding road and rail infrastructure investment, transit funding, zoning and land use planning, public services, etc. And with a lag time of 5-10 years or so. (At least - today's private decisions about where to live and work, and how to get there, are made based on infrastructure and planning decisions made decades ago.)

So it's just really not helpful or interesting to make cross-country comparisons based on raw commute times.

I'm not even going to touch the Cato Institute(!!!) report on Portland. (I've lived there, I know urban planners there - it's not paradise, public transit in particular could be a lot better, but the growth planning has been largely a success on the appropriate metrics.)

Posted by jack lecou | May 7, 2008 1:46 PM