It almost goes without saying that congestion is a terrible thing, so bad that it justifies the spending of massive amounts of public money as well as the impact on our cities from widening and building new transport infrastructure to rid ourselves – or at least reduce the level of – this terrible thing that is congestion. So you would expect cities with lots of congestion to be horrible places that are struggling to attract population and have a poor quality of life – while you might expect cities with less congestion to be great places that are attracting heaps of people and have a great quality of life. Right?

Well the reality appears to be quite different, as touched upon in this recent Planetizen article – which compares cities in the USA with some of the highest levels of congestion with those that have some of the lowest. Let’s start with the more congested cities.

Wendell Cox just wrote an essay trying to correlate density and congestion, asserting that density means congestion and congestion is really, really bad (or in his words, “less traffic congestion benefits a metropolitan area’s competitiveness.”)

So logically, the high-congestion cities should be declining, and the low-congestion places should be attracting Americans at a rapid rate. Right? Wrong.

In fact, the lowest-congestion cities tend to be a very mixed bag, while the high-congestion cities are doing relatively well. Cox lists ten high-congestion regions: Los Angeles, Houston, Austin, San Francisco, New York, Seattle, San Jose, Washington, Boston, and Portland. In all ten, the central city of the relevant region gained population between 2000 and 2010. These cities tend to be larger, relatively wealthy, high-cost cities, cities where keeping housing affordable is a bigger problem than demolition of worthless vacant lots.

And in all but two of these ten regions (all excepting Boston and Washington) the central city is more populous than in 1970. In these regions, there’s enough growth for city and suburb alike. Although some of these regions experienced regional population growth of 0-10 percent, not one of them shrunk, and two (Houston and Austin) grew by over 20 percent.

Reeling off cities like Houston, New York, Seattle, Washington, San Francisco and others hardly appears to be a list of US cities which are doing particularly badly at the moment. Even though they are apparently the most congested cities. Now let’s look at the least congested cities:

By contrast, Cox lists ten low-congestion regions: Indianapolis, Oklahoma City, Salt Lake City, St. Louis, Richmond, Kansas City, Memphis, Buffalo, Rochester, and Cleveland. A few of these (most notably Indianapolis, Oklahoma City and Salt Lake City) are doing reasonably well. But five of the central cities in Cox’s “hero metros” lost population in the 2000s (Buffalo, Rochester, Cleveland, St. Louis, and Memphis) and two more gained population in the 2000s but are still less populous than in 1970 (Richmond and Kansas City). In fact, Buffalo and Cleveland even managed to lose population regionwide, and not one of Cox’s high performers grew by more than 16.7% (metro Salt Lake City’s growth rate).

In addition, these low-congestion cities tend to be far more dangerous than high-congestion cities. Their average murder rate in 2012 was 19.5 per 100,000 residents, while the high-congestion cities’ murder rate was only 7 per 100,000—not surprising given the decline discussed above. Only one of the low congestion cities (Salt Lake City) had a murder rate as low as the average for the ten high-congestion cities.

Residents of Cox’s ten low-congestion cities have more reason to worry about dangerous drivers as well as dangerous criminals. I was able to find data for auto-related fatalities for sixteen of the twenty cities in Cox’s two “top ten” lists; the high-congestion cities averaged 4.7 traffic deaths per 100,000 residents in 2011, while the low-congestion cities averaged 9.8. To put the matter another way, the most dangerous of the high-congestion cities (Houston) had 9.1 traffic deaths per 100,000 people, while five of nine low-congestion cities had more traffic deaths per capita than Houston. The second most dangerous high-congestion city (Austin, clocking in at 6.4 deaths per 100,000) had a lower fatality rate than all but one of the low-congestion cities. The least dangerous of the low-congestion cities, Rochester, New York had a higher traffic death rate than five of the seven high-congestion cities.

St Louis, Buffalo, Cleveland etc. at first glance appear to be some of the US cities that have most struggled over the past few decades – losing a heap of population as US manufacturing moved off-shore. The very low congestion levels enjoyed by these places doesn’t seem to have any effect on their relative attractiveness – and seems to be linked with higher murder rates and much greater risk of traffic related deaths.

Michael Lewyn, the article’s author, suggests a possible reason for the connection between low congestion cities and these fairly poor statistics:

Or it could be that policies designed to limit congestion (like widening roads to support high speeds, chopping up downtowns with highways and turning them into giant parking craters) have actually had some positive effect for congestion, but at a heavy cost.

Given that generally we know congested cities are more economically productive, is it time we stopped stressing about this issue so much and focused on things that really matter like levels of accessibility and the extent to which people are able to get around unaffected by the particular road conditions of the moment. Of course that’s a key driver behind the the Congestion Free Network.

CFN 2030A

Share this


  1. Nice work Matt L – take that, petrol-heads out there!

    Of course it calls into question the extent of Auckland’s CBD decline over the 1960s, 70s and 80s. While there were certainly US led global trends at play, it might also be said that the near death of Auckland’s CBD was a consequence of the massive severance effects created by Spaghetti junction and the motorway network, which ironically of course helped traffic flow more smoothly elsewhere in the region, at least for a while.

    More recently, as the CBD started to recover and Auckland has started to take on the role of being a more outward looking international city, supported by gradually improving PT, road congestion across the region has also increased. So – there is connection between congestion and the vibrancy of a city – the debate of course is: causation or coincidence? Do we as a nation want Auckland to be a big provincial town with literally no heart, or an international city?

  2. The issue here is that congestion is *associated* with economically productive regions or cities, not that congestion actually drives economic productivity.

    It doesn’t surprise me that the economic powerhouse cities have higher rates of congestion at all – they’re where people want to live, where people are moving in multiple directions, with large employment centres for the daily commute. It’s a very large stretch to then say that congestion is an objectively *good thing* for cities’ economic performance. I would hypothesise that once you control for the other variables which drive both economic productivity and congestion, congestion emerges with a negative residual impact on economic productivity.

    1. Well said Xavier. Congestion is definitely “a negative residual impact”. Not good but a sign of good things.

    2. If you follow this logic then it seems the “Congestion Free network” is promising to do a bad thing . . . perhaps you should drop the “Free” from the Label ?

      1. Well Tony’s quite right really, it’s more the Congestion Irrelevant Network (Haha: CIN). The network that allows citizens to never bother thinking about congestion again, or at least to opt in only if they feel like it. The name we gave to what is simply the obvious Rapid Transit Network for Auckland reflects the obsession with the idea of congestion in any city that only invests in driving amenity. You get what you pay for: No choice but to join the queue otherwise, ah the freedom of the road…..

  3. Regarding the correllation between low house prices and congestion, and high unemployment and crime rates, I am reminded of a quote from GTA3 character Donald Love: “Nothing drives down real estate prices like a good old fashioned gang war, apart from the outbreak of a plague, but that might going too far in this case.”

  4. I don’t pretend to understand all the issues here, but if cities are succeeding despite congestion (which is measuring stress in one mode), then surely other the transport modes are performing well, or are strong enough that residents still have adequate mobility?

  5. Post hoc ergo propter hoc. Great news! all we need is some road cones and our economy will flourish.

      1. not quite. It was just ridicule. r-a-d is trying to show something is true by showing something false or even absurd follows from its denial. I was just making the point that because something co-moves it doesn’t mean it is caused by it.

  6. All this shows is that there becomes a point when the money you are spending to “solve” congestion has no real return (a point Auckland may have already passed or is rapidly approaching). No more, no less. Having high congestion doesnt mean a city will succeed, but neither will low congestion.

    And of course totally ignores that fact that all the users of grade separated PT, pedestrians and cyclists are not affected by congestion in their journeys.

    That is entirely consistent with everything this blog has said. Trying to state (as mfwic inevitably has) that this post posits a causative connection between congestion and successful cities is not borne out by the actual words in it. It is as ridiculous (well maybe not quite) as claiming low congestion makes a city successful.

  7. I don’t think more congestion is a bad thing in the sense that things are safer for pedestrians or that less congestion is a good thing (like Detroit). It depends entirely of the impact of the congestion. If that congestion only occurs for 1-2hrs of the day when everyone is going/leaving work at the same time, then you have already spent too much money on the problem and ROI will be low. If the roads are gridlock 24-7, then maybe you could do with some more transport spending.

    So in our case, we should be focusing on PT spending, while maintaining the existing private vehicle infrastructure.

  8. In the 60s in the US, many inner city retail districts closed their streets to make pedestrian malls thinking that having more space for people and less – no – space for cars would boost sales. It took about 30 years but it was finally realised that there’s gold in them thar cars, and virtually every urban pedestrian mall was torn out. “Congestion” came back and so did shoppers. It was a hard way to learn a simple lesson.

    Congestion needs to be rebranded. It means that there are a lot of people and they have someplace to go. Some cities don’t have congestion because neither of those factors apply. One problem I have with interpreting the info here is that between the lines there seeks a correlation of congestion/growth/density. If so, then Houston is an outlier. It’s not dense and it’s congested because there are a lot of people and they have someplace to go, but those places are far away. (Hey! Helicopters!) But congestion mean people.

    I don’t have to read the study to know what it’s trying to do. Wendell Cox is a libertarian crank who devotes his life (funded by right-wing “think” tanks, etc.) to killing public transport. (In view of the logical fallacies being pointed out here, I hasten to say that this isn’t an ad hominem attack, just a conclusion that’s easily drawn from familiarity with about two decades of his work. He has some good mate in NZ, too. I don’t remember his name. Owen Somebody?) There are always holes in his analysis and this is just more of the same. Fortunately he has no meaningful influence outside the libertarian cave.

    Anyway, I was saying… he is simply trying to say – for the x thousandth time – that city bad, suburb good. Government (read: PT, minorities, liberals) bad, suburban developers good. It has always been thus with Wendell Cox. So the analysis is a good take down.

  9. Then there is the issue for the high cost of incremental benefits of so called ‘congestion reduction’ spending. The government has just announce 800 million dollars worth of motorway widening in Auckland, with money conjured from a variety of sources: loans, asset sales, and the NLTF.

    These works will no doubt speed some drivers journeys for a while both in space and time. In space, until they get to the next constriction, or onramp adding more drivers or whatever, and in time because this spending incentivises more driving which will inevitably lead to these stretches becoming clogged again. Doubly so because this sum of money is not being spent to improve any alternative to driving for all journeys and at all times therefore preventing any likelihood that people will choose to not drive even for some journeys.

    For anyone seriously grumpy about congestion must be interested in getting people to leave their cars at home, especially at peak times, which surely means improving those alternatives and not encouraging more driving, especially at the margins.

    In truth this 800 million is investing in congestion, buying more of it. It really must be something that the government likes and wishes us to have more of.

    As the Council faces the reality of the rates rise cap at 2.5% over the next couple of weeks I hope it understands that it has a responsibility to try to balance this decision by government, and makes all of its cuts to road projects all across the region.

    In order to balance our transport infrastructure we need to invest asymmetrically in the missing modes. Then congestion will simply be an ‘opt in’ experience.

  10. I’ve always wondered what happens when one mode overloads or breaks down, will other modes in the same region pick up the slack or is the overload contagious?

    I see three major modes for personal transport in the Auckland region:
    – Cars
    – Buses
    – Trains

    There are a number of smaller modes (Walking/Cycling/Ferries) that provide some capacity and support the overall system.

    What I struggle with is why we continue to invest in a single mode, when balancing other modes will provide choice and mutual support to other modes, along with an amount of fault tolerance.

    Each option has its strengths and weaknesses. The article/paper being discussed has been presented as having a bias towards one mode over others. I think that some congestion is healthy, as it will help users make value judgements that optimise resource allocations, predominantly time or cost.

    In Auckland the investment bias towards roading is distorting the benefits that can be achieved by investment in other modes. How long will this continue?

    What would be the result of the shutting of a key piece of infrastructure (bridge or viaduct), how would the system cope ?

    1. Yes Auckland is completely out of balance, and the returns on continuing to keep investing huge sums in the one mature and lavish system get ever more marginal every day. However the three modes are actually:

      Driving: Car, Truck. Motorbike
      Transit: Bus, Train, Ferry
      Active: Walking, Cycling

      The Minister argues that because current use is overwhelmingly on the first mode then it must have ever more sums thrown at it [ie it must be encouraged further].
      This argument is based on the unsupportable assumption that the current ratio is permanent and perfectly expresses everyone’s choice. Why is it unsupportable? Because it constantly changes in response to what options are available. Both Britomart and Northern Busway have been responded to by people gladly using them. Properly joining up and speeding our still feeble Transit system, especially with a true Rapid Transit Network as imagined in the CFN, will undoubtably lead to much much more use.

      Or is this what the Minister is afraid of?

      1. At here is another way of looking at it Patrick. Can one mode ever be truly considered separate from another mode, if they share the same network? After all, it is the medium they travel on, ie the network itself, that clogs and congests, not the car themselves. Therefore the modes are:
        A – everything that travels on a road.
        B – everything that goes by rail.
        C – everything on water,
        D – everything by air.

        Both C and D clog only in the nodes, but aren’t really relevant in Auckland. So the choice comes down to A or B – Road or Rail. So the road network contains not just cars and trucks and vans, but also buses, and, importantly, it also concerns cyclists and pedestrians, who typically have to use the same roading network ie footpaths along side the clogged roads.

        So if you really want a choice of networks by which to seek to avoid congestion, then a completely separate network needs to be built up for buses (the Northern Busway is a good start) and cyclists (long way to go there…). The best way to get people to use another form of transport (ie buses or trains) rather than cars, is to let the cars totally clog up until they stop moving. But only if you have an alternative system already in place. With just the one system in place ie roads, there is no redundancy in the system.

        1. Except buses on busways and buslanes are not subject to congestion and nor are bikes on bike lanes. Furthermore motorways without busways and stations are a single mode investment. So the ‘buses need roads too argument’ doesn’t wash with the latest 800 million. With integrated fares and stations designed to facilitate cross mode journeys trains and ferries do have more in common with buses from a user perspective, especially once there are more buslanes so the buses have their own ‘tracks’ as it were.

          Cars, trucks, and motorbikes however are not stuck with fixed routes.

          So I think my distinction is better.

          Of course there are many mash-ups:
          Every Transit user is a walker too, as are many drivers, in practice, even though the traffic models don’t think so [everyone leaves from their bed and parks by their desk].

      2. Simple really. Roads push as much as possible of the operating burden onto the road users. Travellers have to provide their own vehicles at great expense and hardly any staff are needed to keep the system running because everyone has to be their own driver. It’s a wet dream for tax anorexics and those smug hierarchicalists who love to blame those less fortunate than themselves for not having a car/education/opportunities-to-develop-as-part-of-a-well-adjusted-childhood.

        1. As much as possible but not enough – the majority of local road costs are still paid out of rates and general taxation. So even with a lot of the costs passed onto the user, travelling by car still requires massive subsidies.

          And now that so much money is being sucked up by the RoNS, even more public money (like the government’s election slush fund from assets sales) is now needed to keep even the roading system normally paid for by RUCs and excise tax plodding along.

          Fiscal responsibility?

Leave a Reply

Your email address will not be published. Required fields are marked *