Every weekend we dig into the archives. This post by Matt was originally published in January 2014.

Most proposals to build new roads or widen existing ones seem to boil down to an ultimate belief that it will “help the economy”. Whether it’s by improving freight reliability or getting people to their jobs faster or helping business travel or whatever, there seems to be a fundamental belief among many that quite a strong relationship must exist between building more roads and improving the economy.

Clearly this is a contestable assumption, and some recent research in the USA details some pretty interesting trends – as reported on in Planetizen:

University of Minnesota professor David Levinson has written in the past that, because of the relative completeness of our national highway network and the cost of construction, the return on investment for additional mileage is approaching zero. One study estimates the return on investment for highway construction was just 14% between 1990 and 2000.

I recently decided to follow up on this line of research, so I dug through some Census data. What I found was shocking, though not altogether surprising. It seems that, besides wasting billions of taxpayer dollars, road-building may actually be holding back economic growth overall: from roughly 2000 to 2010, states that built the fewest urban road miles grew an average of 64 to 94 percent faster than their asphalt-enamored neighbors. Rather than increasing productivity through increased mobility and reduced congestion, as politicians and lobbyists so often promise, all this mindless road-building could be depressing statewide economic growth!

Let’s look at the details a bit more:

Looking at the numbers in aggregate, we see some interesting trends that seem to hold up just about any way you slice the pie:

  • States that increased their urban road mileage by less than 30% grew by an average of 14.40%, while those that increased mileage by greater than 30% grew by an average of just 8.77%.
  • If we set the cutoff at 20% mileage growth, states that built less grew by 17.97%, and states that built more grew by 9.24%.
  • At a 10% cutoff, states that built less grew by an impressive 20.70%, compared to just 10.66% for those that built more.

Statistically, analyzing the correlation between road-building and economic growth gives us an r-score (correlation coefficient) of -0.34, which implies that about 10% of a given state’s economic growth can be explained by how much urban road-building they did over this time period. Many things influence the overall health of any economy, obviously, so we shouldn’t expect the quantity of roads to wholly predict statewide economic growth by itself, but this does indicate a negative correlation between the two variables: more roads equals less growth. (As always, please remember that correlation does not imply causation.)

And for a graphed comparison:

roads_vs_growthThe post’s author, Shane Phillips, doesn’t think that these results are particularly surprising:

None of this should be particularly surprising. While politicians and advocates love to tout the job-creating value of new road and highway capacity, congestion reduction rarely lasts more than five years and widened roads ultimately only succeed in extending the boundaries of wasteful, unproductive sprawl. In the case of road widenings, it’s entirely possible that the disruption caused during the construction phase completely erases — or even exceeds — the fleeting benefits of reduced congestion.

Then there’s the opportunity cost: think of all the good that could have been done with the hundreds of billions of dollars spent on roadways over that period: more responsible transportation spending, education, renewable energy … take your pick.

I think it’s probably unlikely that building roads directly harms the economy, but there are logical reasons to think that it might cause indirect harm: particularly due to it not the best use of public funds and encouraging dispersed land-use patterns which undermine agglomeration. New Zealand’s heavy dependency on private vehicles also forces us to spend a lot of money each year importing cars and oil – basically cancelling out wealth that we create from exporting dairy to the the world.

The next version of the Government Policy Statement will be released some time later this year. If it’s anything like the current version it will stress the importance of transport’s role in improving the economy and then make a giant leap of faith in assuming that building more roads is the best way for transport to improve the economy. It’s time to fundamentally question that assumption.

Share this


  1. You can only use least squares regression when you have a dependent variable and at least on independent variable. When you have mutual causality the maths simply fails. It gives a result but the result is completely without meaning. In this case it is possible that states that have higher economic growth are those that can afford to build metro systems. So growth impacts on road building and road building impacts on growth. Least squares has numerous corrections for different types of bias but there isn’t a correct for this type of endogeneity. You have to use a proxy for one variable or else solve as simultaneous equations.

    1. “In this case it is possible that states that have higher economic growth are those that can afford to build metro systems.”

      If that’s the case there are even bigger problems because the cost to the city of building roads is higher than that of developing good public transport.

      1. While the maths doesn’t work the plot does. If you look at the upper left hand part you will see four outliers that are leading to their conclusion. The top point is not a state but Washington DC where the economy grew quickly due to vast increasing in the security budget.
        Below that at 28% growth is North Dakota which had an oil boom and to the right and up above 30% are Wyoming (Minerals, Energy and tourism) and Louisiana (Oil and rebuilding after a hurricane).
        This is almost as bad as the dude who concluded sun spots were the cause of economic cycles.

  2. “I think it’s probably unlikely that building roads directly harms the economy”

    I suppose the environmental cleanup of the damaged waterways, and the medical profession employed in treating the people affected by transport system-induced physical inactivity helps the money go round? :/

    The carbon credits we’ll be buying to mitigate our increased transport carbon emissions are one direct way that road building directly harms the economy. And at an international level, the transport carbon emissions from road building will cost countries dearly.

    But when you wrote this article in 2014, It wasn’t PC to discuss these concepts. We’ve come a long way in 5 years.

    1. Heidi, I am not sure whether it was PC to talk of climate change in 2014, but Council certainly released the Low Carbon Strategic Action Plan that was almost immediately consigned to history.
      It is a matter of record that no progress has been made by Council toward the then target of reducing emissions by 40% by 2040. (I wonder because of largely universal global inaction the target is now 45% by 2030; and that in a handful of years time it might be 50-55% by 2030).
      I think we need a time when it is not PC to talk about road building, more car parks and the silly notion that getting freight from say Auckland to Whangerei 10 minutes faster is somehow going to be hugely stimulatory for the economy.

    2. I’m not sure PC is the usual term. It seems you’re talking more about something like “the Overton window”:

      the window contains the range of policies that a politician can recommend without appearing too extreme to gain or keep public office in the current climate of public opinion.

      via Wikipedia. https://en.wikipedia.org/wiki/Overton_window

      I mean, being opposed to climate change has been a politically incorrect opinion for years, but that hasn’t previously translated into using climate change to justify non-pollution/carbon/etc policies. On the other hand, the willingness to consider roads as a problem might not be related to increased acceptance of roads-as-global-warming but instead a kind of creeping normalisation of pro-PT policies. That those policies also have climate change related/environmentalist justifications is helpful.

      For example, I’ve heard (via an IRL acquaintance) the Herald decided not to publish an opinion piece to the effect of “If the CRL over-runs are due to high demand, why not hold off on the next stages until the big infrastructure market’s prices have cooled?” It’s only six or so years ago that the Herald was perfectly willing to put opinions like, “I’m a law grad and I think the CRL needs more business cases since it’s based on the notion of ‘if you build it, they will come’ which is an absurd claim on its face” on its website. I know the CRL isn’t quite the same kind of thing as what we’re talking about but it’s an interesting shift. (Especially given the former is actually the more intellectually credible case.)

      In fact, I’m not really so convinced we are ready for cars-as-global-warming even if people will now believe the-only-reason-NZers-don’t-catch-trains-is-that-they-didn’t-have-any-to-catch. The former is an idea a few brave souls will put out there to like-minded thinkers, but I don’t think we’re in an early adoption phase per se. I think it’s still more about adding environmentalism to policies/ideas still conceptualised as being about congestion, economic efficiency, equity or whatever.

      1. I was going to use ‘Overton Window’ but do have fun stirring by using the term PC instead. 🙂 It originally meant being sure to say things that were socially acceptable in the (then) political climate. To keep the term close to the meaning of the words that make up the term, that’s what it should still mean. The particular political climate when the term was coined means it came to mean a particular way of being correct – avoiding terms or statements that could be offensive. I find it’s a far more useful term to return it to its roots, and comment on whether someone is saying something that fits with the political climate.

        In 2014, the government was ignoring climate change, and it was not socially acceptable in most circles to have an honest discussion about climate change and its ramifications. Certainly policy change and behaviour change were generally no-go areas. (Not that this stopped me, of course).

        Using the original meaning of the term PC, therefore, I don’t believe it was ‘PC’ to do more than use the term ‘climate change’ in passing.

  3. Brexit or no Brexit, dangers remain. Angola plans to join Rwanda and Mozambique (combined population 74 million) in the “Commonwealth”, within a “new network of partners”. Having announced their planet-saving decision to have no more than two children, the tax-funded celebrities known as Mrs and Mr Markle are jetting off to Angola as “fantastic” FO ambassadors to the “extremely young” populations in the “vibrant” continent of Africa (Sunday Telegraph, 15 September).

    1. Seriously? WTF is the Commonwealth for if they let in Rwanda which was ruled by Belgium and then Angola which was Portugese or something. They should tell them to bugger off and form their own ridiculous vestige of colonialism.

  4. Subway or Rail Transit aloud you to rest and read during the daily job travel.
    You don’t need to breath road air during the travel and can afford be aware of parking problems

  5. The opportunity cost isn’t just about the amount of money being spent. There are only so many companies that can finance big projects and they can only finance so many projects at the same time. Whether that is rebuilding a school such is going to happen in Ashburton or building a motorway extenstion as is happening between Christchurch and Rolleston.
    Then you take that to the next level. There are only so many construction companies that can complete these large scale projects. It is harder to scale up a company to take on more projects than it is to scale down after it is finished. No problem if you finish one and start the next much of the work force can be rolled over onto the new job. More of an issue if you start a new project while much of your work force is still tied up in the first one.

Leave a Reply

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