Do cycle lanes actually increase neighbouring trade?

This was meant to be (and someday will be) a different piece entirely. I was reading Simon Kuper's excellent summary of thoughts on the 21st Century City, and thinking about a response. Something, being of an irritating-small-child-tugging-on-a-dress-asking-why state of mind, about the dangers of groupthink and whether such unanimity around the elements making up the modern city was a good thing.

Within Kuper's piece there's some fairly standard anti-car rhetoric, this time in favour of e-bikes. And my mind wandered to the subject of installing cycle lanes, and increased pedestrianization. And I started to think about that phrase which is used by cycling activists over and over again as an economic argument for the installation of cycle lanes. You'll have seen it. It says something along the lines of...

Cyclists spend 40% more in shops than motorists

It's a line that, in various forms, pops up all over the place (not often attached specifically to London as it is here). Especially in council and public sector documents arguing for more cycling and more infrastructure. Google has a LOT of examples; almost every document looking at the economic impact of cycling references it, or something very similar.

But is it true?

Reader, I think you already know the answer...

Tl:DR:

• There's little robust data about the direct economic impact of cycle schemes.
• There is, however, a lot of data pointing to positive impacts from public realm improvements which can include cycle lanes. But...it seems to be walking and pedestrian improvements that make a real difference.
• Where claims are made about cyclists spending more, it tends to be in areas which already have high bike use--like student suburbs in the US--and which are far from predictive about installation of the lanes elsewhere.
• Using data incorrectly, or reaching from one specific situation to an entirely different one, because you *want* something to be true, isn't good policy.

The origin of the 40% stat is a Transport for London document from 2018, Walking and Cycling: the economic benefits. A high percentage of google links to the basic stat (40% increased spend) lead to the press releases TfL put out to accompany the report. It was heavily promoted. And on page 5 it clearly has the relevant data. Or not.

Okay, so public realm improvements (that include cycling improvements) *can* increase retail sales by *up to* 30%.

Right underneath is another stat

Okay, so the 40% is actually people who *walk* to the High Street, and not about cycling at all. For those covering the research paper, especially those with agendas, there's been a dangerous conflation of the two. And note the 'up to' again; that could be 1% of course.

Right, but the first graphic refers to 'Lawlor, 2013' from which the data comes. Perhaps *this* document contains the killer detail. Well, no.

Lawlor, 2013, is a report entitled The Pedestrian Pound, done for 'Just economics for Living Streets.' It is, as you would imagine, over-whelmingly focused on the economic impact of walking, not cycling. It's a really detailed report and is well-worth reading. It contains this very relevant statement (p.19):

Identifying a fully attributable, causal link between investment in the public realm and direct commercial returns is a challenge.

But it tries. Two relevant quotes for our purposes (remember, we're trying to track down the '30%' claim referenced by TfL)

"Nevertheless, well-planned improvements to public spaces within town and city centres have been shown to boost commercial trading by up to 30%" (no reference, p.23)

"Litman estimates that walking and other non-motorised transport projects typically increase retail sales by 30% (Litman, 2002; Burden and Litman, 2011)" (my emphasis, p.23)

Lawlor then goes into a detailed set of case studies that mostly look at increased footfall, rather than spend. But it does mention work done in Brooklyn in 2013, where businesses saw a 48% increase in retail sales after pedestrian and other improvements (see below). There's a reference to a piece of work on Toronto (Tolley, 2011) as well which we'll look at below.

So, Lawlor is not talking about cycle lanes, and though there are a couple of examples of a 30% uptick (and more in the case of Brooklyn) in local business performance, when those figures are mentioned it's in the context of general public realm improvements.

That Brooklyn scheme is a perfect example of the issue. This saw huge sales increase--maybe it's from here that the spend 40% more comes. But looking at the actual scheme in the NYCDoT handbook you can understand why there was such an increase...its's not just about cycle lanes--far from it. It's about trees and pavements and cleaning the whole place up. Interestingly, the rest of the handbook is full of examples of schemes where retail spend increased as much, if not more, thanks to improvements to road crossings, pavements, etc. It's the public realm changes that are important, not the cycling.

So--public realm improvements that might (or might not) include cycle lanes can lead to increases in spend locally. Which is not really a major surprise. But it's not evidence of the impact of cycling itself.

So is there some elsewhere?

(take a breath, there's a deep dive coming)

The other quote up top about the impact of walking referred back to a TfL report of 2013. This appears to be a report on town centres, and is about shopping (among other things) in local town centres in the suburbs of London. This has that 40% figure (p. 55), but clearly for walking.

The total average spend per month by mode (see Table 27) shows that average spend for those who walk to the area was ￡346, for bus it was ￡292, for car it was ￡247, for train/Tube it was ￡192 and for cycle it was ￡190.

Which is not surprising, given it focused mostly on people living in local areas and their relationship to their main shopping streets. The cycling figure, though, is an obvious concern if trying to prove cyclists spend more (and this is taking frequency of trips into account).

Then there's The Value of Cycling (Raje and Saffrey, nd) which contains two economic claims, alongside a lot of others:

• Cyclists visit local shops more regularly, spending more than users of most other modes of transport
• Public realm improvements, including those that cater for cycling, have been shown to result in increased trade at local businesses; up to 49% in New York City (that Brooklyn study again)

The first claim quickly leads to another three sets of references ("Numerous studies have shown that while cyclists spend less per visit, they tend to visit shops more regularly resulting in higher weekly spends") including Clifton, below.

The second refers to a number of studies about design, none of which seem to have robust spend data. As well as the Brooklyn project.

Raje and Saffey say overall,

Traditional cost-benef t analyses of cycling infrastructure do not generally consider the wider/indirect economic impacts, including impacts on local business and retail establishments. The very limited evidence that does exist is generally favourable for cycling infrastructure.

Although they provide no original evidence to support that latter claim.

So the Value of Cycling leads in turn to Examining Consumer Behaviours and Travel Choices (Clifton et al, 2013) which refers to work in the lovely Oregon city of Portland. This contains analysis of consumers at 89 businesses in the city. It, indeed, finds that cyclists (and pedestrians and public transport users) spend more than drivers. But...(and it's a fairly hefty but) its methodology begs the question of how transferable its findings are. It focuses on four sectors (fast food restaurants, bars/cafes, convenience stores (7-11s) and suppermarkets) and interrogates behaviour between 5pm and 7pm Mon-Thursday for the first three and over the entirety of a single week for the latter. These are workers on their way home, picking up a small shop, or stopping for a quick drink before heading home (the prevalence of people visiting bars while on bikes is actually slightly worrying). It's not about destination retail, or clothes shopping, or going out for a meal with the family. It's interesting, but not relevant to planning for city centres. Indeed, in the supermarket sector, it's clear drivers spend most, which is hardly surprising.

(there's an interesting nugget on p.39 in the conclusion though: "Much of the bicycle-specific infrastructure that past literature has shown to be associated with greater levels of bicycling, such as bike lanes and paths, are surprisingly not significant predictors of cycling in our analysis" which clearly puzzled the authors).

It does, however, in turn refer to Tolley (2011), also mentioned above, which is a report for the Australian Heart Foundation. And that in turn refers to analysis of the proposed Bloor Street cycleway in Toronto (Sztabinski, 2009). This states

Patrons arriving by foot and bicycle visit the most often and spend the most money per month

which seems pretty conclusive. The majority of the paper is around the verbal responses of business owners and pedestrians to proposals for an increase in cycle access (most are positive and see it as likely to increase takings). Bloor Street is a very high student-population area, and students are much more likely to have bikes than cars. And merchants who serve students know that, and have set up to serve them. There's a circular argument here--this is how people already live in this (and other student) areas. It's not predictive of what happens when new infrastructure is introduced. Indeed, given that it's pedestrians that are interviewed (which includes some people who have driven to the area and are now walking about, of course), it's no suprise that pedestrian spend massively dwarfs spend by users of any other mode.

As I was writing this, a dear friend (and cycling ambassador) sent me a link to a great Bloomberg piece from 2015 which collates 'Every Study ever conducted on converting street parking to bike lanes'. This focuses on many of the examples already given and some newer ones. It's mainly interested in arguing that converting parking spaces to cycle lanes or cycle parking is effective. There's a nifty study of spending habits on Grafton and Henry Street in Dublin which removes statistical outliers (ie big spenders) using cars in order to demonstrate that 'spending by bicycle users is as high as by motorists. ' Well, if you take out some of the latter's spend data, it would be. Analysis of removal of a small number of parking spaces in Seattle suggests no impact on overall spend, which again seems likely given the marginal contribution those parking spaces would have on overall sales figures. But would conversion of those spaces to wider walking lanes, have contributed as much if not more than conversion to cycle lanes?

All of these studies end up in the same place. There's really just one study (the Bloor Street one) that shows higher spend by cyclists than by motorists and it seems to have taken on an impact way greater than its methodology or specificity deserves. Even this, if we're still loking for that rogue '40%' figure, doesn't get close. So what we have is campaigning for increased cycle infrastructure based on the recroded impact of general public realm improvements and claims about cyclist spend that aren't in any way justified by the research (at least as far as a couple of hours' work suggests).

Now, none of this is a reason NOT to install cycle lanes, of course. There are all sorts of benefits. Just not a proveable economic one. And if it's not proveable, perhaps driving a cycle lane through a high footfall retail or hospitality area, or replacing cars with bikes in a city centre, might have some unintended consequences. Like reduced trade. And increased anti-social behaviour. But I'll return to that point in a future ramble.