Good cycling infrastructure is extremely important for increasing cycling and for reducing the use of cars in our towns and cities. On main roads, that means protected (segregated) cycle lanes; on side roads it means filtered permeability – pedestrians and cycles go through, cars and trucks don’t. There’s plenty of evidence for the importance of such infrastructure.

But when the UK cycling organization road.cc runs the headline “Cycle infrastructure responsible for 85% of cycling increase“, I have to object. This is the message of Infrastructure, Infrastructure and Infrastructure gone mad. If you read just that headline, you’ll be left with the impression that nothing much matters other than infrastructure. Even if you read the whole article, you won’t know why that interpretation is dead wrong.

The road.cc article, by Laura Laker, reports on a study by Prins et al (non-paywalled version here), on the effect of a new cycle path linking the Cambridge (UK) city centre, Cambridge Science Park, Cambridge Biomedical Campus, and villages to the northwest of Cambridge. The people studied all work near the path; what the study found was that the increase in cycling, from before the path was built to after, could mostly be explained by how close to the path you live. That’s the 85% bit. Prins et al. also test certain psychological theories to the effect that the importance of living near the cycle path may itself vary depending on the individual’s perceptions of safety, and attitudes toward driving. But that’s it – the reason the proximity to the new cycle path explains 85% of the change in cycling is that it is the only change that was studied; perhaps it was also the only relevant change in Cambridge during the years studied (2009-2012).

Much of what we know about the importance of cycling infrastructure comes from the Netherlands. A few years ago the London Cycling Campaign adopted the slogan “Go Dutch” – to promote demand better cycle infrastructure. In the past two years, three of London’s outer boroughs have been making themselves into “Mini-Hollands” – by building cycle tracks, getting rid of through motor traffic on side-streets, and so on. So one could be forgiven for imagining that this finding from Cambridge reinforces what we know from the Dutch experience, namely that only infrastructure matters.

But that isn’t what the evidence from the Netherlands tells us. While Dutch studies have found that the proportion of trips taken by bicycle is strongly affected by infrastructure, the same studies show that by making it more costly to drive also makes an important contribution. “More costly to drive” means, in these studies, “more costly to park”, because the price of parking is the policy tool that Dutch cities have to affect the cost of driving.

Comparing 103 Dutch cities, Rietveld and Daniel (2004) found that both better infrastructure, and the higher prices for parking cars, encourage cycling; they estimate that a rise in the price of parking by 0.14 Euros per hour increases cycling share by 5 percentage points. Given the difference between the average price of parking in their dataset (0.17 Euros per hour) and the maximum (1.23, for Amsterdam), this implies that raising a city’s parking tariffs from the national average to the Amsterdam rate would increase the city’s cycling share by a huge 30 percentage points – almost doubling it. That’s just the estimate from one study, so it needs to be taken with a grain of salt, but if it’s anything near correct the effect of pricing parking can be huge.

A more recent Dutch study, by Harms, Bertolini and te Brömmelstroet (2016), reinforces this view. It considers only 22 Dutch cites, but includes information on changes in each of them between 2000 and 2013. Unlike Rietveld and Daniel, the new study doesn’t produce quantitative estimates, just rough classifications of what policies work under what circumstances; the lack of quantitative estimates is a limitation, but their method allows them to consider combinations of factors in a way that Rietveld and Daniel could not. Harms et al find that in cities with low cycling shares (that’s low for the Netherlands, which is still high for the rest of us), the cycling share rises in response to a combination of infrastructure improvements and increased cost of parking.

The Cambridge study is a good study of the effect of a single big change (a strategically positioned cycle path) in a setting where little else changes; under those circumstances, however, it is bound to find that most of the effect (that 85%, again) is the result of the one thing that changed. It tells you absolutely nothing about the relative importance of different factors in promoting a shift from driving to cycling. The Dutch studies, by comparing a large number of cities, are able to consider the impact not only of the quality of infrastructure and the cost of parking, but such things as topography (hills), city size, the age and income of the population, political leadership, provision cycle training, and so on. The take-home message from them is that both infrastructure and the cost of driving make important contributions to the cycling share. Think of those as pull and push factors: good infrastructure makes cycling more attractive, costly parking makes driving less attractive. Political leadership on the issue also helps, while training people to ride bikes safely makes no significant difference – findings that will not surprise cycle campaigners.

If you were only worried about your own ability to cycle safely where you live, then a single-minded focus on infrastructure might still make sense for you. But if your interest is in boosting cycling as a mode of transport for others – maybe due to concerns about urban air pollution, or about greenhouse gasses (which even “green” motor vehicles will produce until the last watt of electric power is carbon free), or for the public health benefits of less sedentary lives, or because you want an urban environment in which it is safe and pleasant to walk, with thriving high street/downtown shops/stores(*) (UK/US: divided by a common language…) – then you should attend, too, to the need for policies which raise the cost of driving.

(*) If it surprises you to hear that low-cost parking hurts small shops, consider that ample free parking is mostly a superstore / shopping mall thing. Put a price on parking that reflects the damage those shopping trips are doing to the environment, and the car-centred superstore business model suddenly looks less appealing – to the benefit of shops you can reach by foot, bicycle, or transit. See my paper Small, Local and Cheap, for details: in Spatial Economic Analysis, or the non-paywalled working paper version here.