The economics of bike lanes

Montevideo knows what a bike lane costs. The 18 de Julio bike lane, four kilometres long and named for the capital's main avenue, came to approximately 49 million Uruguayan pesos, by the account the Intendencia de Montevideo — the departmental government, which runs the city — published when the lane opened in December 2023.¹ At the exchange rate of the time, that was the equivalent of roughly US$1.2 million. The 4.7 kilometres on the Rambla Sur, the southern stretch of the avenue that follows the city's waterfront, opened in September 2024 and cost close to 68 million pesos, about US$1.7 million. That is the figure the authorities gave the Junta Departamental, the department's legislature, as the news outlet Subrayado reported it.²

What a lane gives back is a different kind of number. For Montevideo's lanes, no such number is known to have been published. For other cities there are several: researchers and city governments in North America, Europe and Latin America have counted customers, read sales-tax filings and built benefit-cost models. What they found is more modest than a boom, and more reassuring than the fear that a lane kills trade.

In outline: most studies that measured trade found that a bike lane did not hurt it, and the three studies that found losses overall did not test their figures statistically. Gains were common, but the largest reported figure shrinks once comparison streets are counted, and a rise in sales on a street with a new lane is not yet evidence that the lane produced it. Some businesses that depend on customers arriving by car can lose. The firmer economic case lies elsewhere — in jobs, in health, and in benefit-cost studies that now include several Latin American cities.

The customers who do not arrive by car

When a bike lane takes parking from a shopping street, the cost to the shops depends on how many of their customers arrive by car. Two studies asked.

In Berlin, von Schneidemesser and Betzien surveyed 145 traders and 2,021 shoppers on two shopping streets, Kottbusser Damm and Hermannstrasse, between June and September 2020. The traders estimated that 21.6% of their customers came by car. The share measured among the shoppers was 6.6%.³ On Bloor Street in Toronto, fewer than 10% of customers arrived by car.⁴

The Berlin traders, in other words, put the figure at more than three times what it was. A shopkeeper who overestimates the car by that much is likely to overestimate what a lost row of parking puts at risk.

Toronto: more customers, and no sales data

In 2016 Toronto removed 136 on-street parking spaces from a stretch of Bloor Street to make room for a pilot bike lane. That was about half the on-street parking on the stretch, and about 10% of the convenient customer parking once nearby car parks are counted.⁴ Arancibia and six co-authors studied the street before and after the change, with a control: Danforth Avenue, which had no bike lane. Their paper, published in the Journal of the American Planning Association in 2019, rests on 3,005 surveys of visitors and on surveys of merchants, 452 on Bloor and 73 on Danforth.

On Bloor, the number of customers rose, and so did the monthly spending those customers reported. The share of all visitors arriving by bike went from 6.9% to 18.1%. Two things keep this from being an account of a lane lifting trade. The study had no sales data: its figures are what visitors and merchants said, not what the tills recorded. And Danforth, the street without the lane, showed similar growth. The authors describe their result as “consistent with an improving economic environment at the intervention site”.⁴ That is evidence of no harm in what customers and merchants reported. It is not evidence that the lane brought the customers.

New York: 49% against 26%

In 2012 New York's Department of Transportation reported an increase of “up to 49%” in retail sales among locally-based businesses on Ninth Avenue between 23rd and 31st Streets, “compared to 3% borough-wide”.⁵ The department's fuller report, in 2013, set that stretch beside three nearby avenues, where sales had risen by 26% on average.⁶

The detail is in the 2013 report. The lane, protected from traffic by a row of parked cars, was built together with left-turn lanes between 31 July and 30 October 2008. The measure was sales-tax filings, set against the quarterly baseline from before the works. Sales among locally-based businesses on Ninth Avenue stood 17% above that baseline in the first year, 47% in the second and 49% in the third; the comparison avenues averaged 26% above theirs in that third year, with Eighth Avenue at 13%, Seventh at 20% and Tenth at 44%. Manhattan as a whole stood 5% above its baseline in the first year, 7% below it in the second and 3% above it in the third.⁶

“Locally-based” has a precise meaning in the report: businesses whose sales-tax filings give a storefront address inside the study area. That leaves out many chain stores, so the sample, in the department's words, “tend[s] to be locally-based mom-and-pop stores and independently operated franchises”. The department also notes that the number of businesses reporting sales rose, “suggesting that new businesses may have entered the area”.⁶

So the fair comparison is not 49 against 3. It is 49% on Ninth Avenue against 26% on the avenues around it — still a wide margin, in a part of Manhattan where sales were climbing on the neighbouring avenues too, and where one of them reached 44%. The figures show a street that did well after its lane was built. They do not isolate the lane as the reason.

Who loses, and what a later study found

A review by Volker and Handy, published in Transport Reviews in 2021, puts the single cases in proportion. It covers 23 studies from the United States and Canada, 15 of them on the effects of particular facilities. Of 35 bicycle facilities, the effect on local business was positive for 20 (57%), unclear or absent for 10 (29%) and negative for 5 (14%).⁷ None of the three studies that reported negative effects tested its results statistically. The six studies that did found effects that were positive or not significant.

The review also names the businesses for which the news is worse. Its more robust studies, it says, suggest that bike facilities “could have negative economic effects on auto-centric businesses”: petrol stations, car repair shops, car-parts dealers and large home-goods stores.⁷ A street can come out level or ahead while a repair shop on it comes out behind.

Vancouver shows what a reported loss looks like. In July 2011 Stantec Consulting delivered a study, prepared for the City of Vancouver and four business organisations, of two separated bike lanes in the city centre, on Hornby and Dunsmuir. Businesses were asked how their sales had changed over the year; on Hornby the change they reported was 10 points worse than on the comparison street, and on Dunsmuir 4 points worse — by the study's estimate, 2.4 million Canadian dollars in lost sales over a year.

Those hit hardest depended on convenient parking. The study ran for two months and relied on what businesses reported. Little detailed sales data came in, it notes, and what did arrive indicated a loss smaller than the one reported. It called itself “preliminary and short-term”, judged the impact moderate and expected it to diminish over time “as long as mitigation strategies take effect”.⁸ It is one of the three negative studies in the review, the ones without statistical testing.⁷

A later case is Chicago's. The city's Department of Transportation compared six commercial streets that have bike lanes with nearby control streets that do not, following sales-tax revenue, employment and vacancies among other indicators; its case studies are dated January 2026 and caution that “it is difficult to distinguish causation and correlation between bike lanes and economic activity”.⁹ As Streetsblog Chicago, on 19 May 2026, and Block Club Chicago, on 23 June, reported the study, the lanes did not hurt business activity, and the study does not claim that the lanes themselves improved any area's economic outlook.¹⁰¹¹

What customers on bikes spend

A study from Portland State University offers a mechanism. Clifton and colleagues surveyed customers at 89 businesses in the Portland area.¹² At three kinds of business, customers who came by bike spent more per month than those who came by car: US$81.76 against US$68.95 at convenience stores, US$81.90 against US$40.78 at bars, and US$48.40 against US$40.06 at restaurants.¹³ The cyclists in those samples were few — 19, 20 and 29 people, against 119, 41 and 174 drivers — so the amounts are indicative.

Supermarkets were the exception: there, cyclists did not out-spend drivers. And per trip, once demographics were controlled for, customers who came by bike and customers who came by car showed no statistically significant difference in spending at convenience stores, bars and restaurants. The monthly difference came from frequency: the customers on bikes came more often.¹²¹³ What the study supports is narrower than a claim that cyclists are better customers: a customer who arrives without a car is still a customer and, at some kinds of business, a more regular one.

Where the case is firmer

The shop till is a narrow place to look for the return on a bike lane. The larger sums are counted elsewhere: in jobs, in health and — with heavy reservations — in property.

On jobs, the reference is a study from the United States. Heidi Garrett-Peltier, of the Political Economy Research Institute at the University of Massachusetts Amherst, examined 58 projects in 11 US cities and estimated, in June 2011, that bicycle infrastructure creates 11.4 jobs for every million dollars spent, counted within the state where the money is spent, against 7.8 for road-only projects.¹⁴ The estimate comes from an input-output model of the US economy of that time. What travels is the direction — a dollar spent on cycling projects buys more labour than a dollar spent on roads alone — not the figures themselves.

On health, there is a peer-reviewed benefit-cost study from Canada. Whitehurst, DeVries, Fuller and Winters, in a paper published in PLOS ONE in 2021, weighed planned investment in bicycle infrastructure in Victoria, Kelowna and Halifax against two benefits only: premature deaths prevented and carbon emissions reduced. The benefits came to between 1.7 and 2.1 times the cost under a moderate scenario, in which cycling's share of trips rises by two percentage points, and between 3.9 and 4.9 times under a major one, in which it rises by five.¹⁵ Retail trade and property values play no part in those ratios.

Copenhagen puts a price on each kilometre. The city's Bicycle Account for 2010 gives a net social gain of 1.22 Danish kroner for every kilometre cycled and a net social loss of 0.69 kroner for every kilometre driven by car.¹⁶ The account says the sum includes factors such as transport costs, security, comfort, branding and tourism, transport times and health, and calls health the most important of them. It does not say that the cost of building the lanes is subtracted. The 2022 account gives a gain of 7.11 kroner per kilometre by bicycle and a loss of 1.26 by car, calculated with the Danish transport ministry's 2022 model.¹⁷ That is a different method, and the 2022 figures cannot be set beside those of 2010 as if the gain had grown.

Property is the weakest of the three. The figure available comes from Indianapolis, where a 2015 report by Indiana University's Public Policy Institute found that the assessed value of the 1,747 parcels within 500 feet of the Indianapolis Cultural Trail, about 150 metres, rose 148% between 2008 and 2014. The report carries its own caveat: part of the increase is a return to earlier values and “likely would have occurred regardless of the Trail construction”, and 25 properties account for 68% of it.¹⁸ The trail is also far more than a bike lane: a protected path for cycling and walking, about 13 kilometres long, with US$4 million of public art along it.¹⁹ The figure describes what happened around a large public project. It cannot be read as the effect of a lane.

Latin America has run the numbers, mostly as projections

Benefit-cost studies of bike lanes are not confined to North America and Europe. Five Latin American cities have figures, and they are not the same kind of figure. One was calculated after the lanes were built. Three are projections. One is a model's estimate for an existing network.

The one calculated after the fact is from Buenos Aires. The Path Less Travelled, a report published in December 2023 by the World Bank and ITDP, a transport nonprofit that promotes cycling policy, looks back at the 17 kilometres of protected lanes built in 2020 on two avenues, Corrientes and Córdoba. According to the report, every dollar invested generated 5.7 dollars in benefits, for an internal rate of return of 113%; health accounts for about 60% of those benefits and time savings for 28%.²⁰ The report does not detail how it arrives at the 5.7.

For São Paulo the figure is a projection. A social cost-benefit analysis of the city's cycle network, prepared by the consultancy Decisio for the World Bank and dated 18 March 2022, describes itself as a quick scan and looks 100 years ahead. Its benefit-cost ratio runs from 6.2, if cycling's share of trips rises from 0.9% to 1.6%, to 30.6, if it reaches 6.7%.²¹ The analysis carries its own warning. In the higher scenario the road-safety item turns negative, as more new cyclists are exposed to traffic.

Lima's is a projection too. The World Bank and ITDP report includes a “what if” analysis of the city's plan for 1,300 kilometres of bicycle infrastructure, which arrives at 19 to 1 on the assumption that cycling's share of trips rises from 0.9% to 15% by 2050. The report itself cautions that its case-study evaluations “cannot be directly compared”.²⁰

For Santiago, SECTRA, the transport-planning body of Chile's transport ministry, presented a social evaluation of the metropolitan cycle-lane master plan in August 2014: an internal rate of return of 31% over 20 years.²² It is a projection, and the figure comes from a slide presentation.

Bogotá's numbers come from a model. ITDP and the FIA Foundation estimated in October 2022 that the city's network of protected lanes cost about US$130 million to build, and that each year it yields about US$80 million in time and money saved by the people who use it and US$230 million from some 300 premature deaths prevented.²³

Montevideo: the cost side of the ledger

In Montevideo the published figures are costs. The 18 de Julio bike lane follows the avenue from Bulevar Artigas to Carlos Quijano and then runs on San José and Florida to Plaza Independencia. The Intendencia gives its length as four kilometres, and its 49 million pesos covered painting, the separators that mark the lane off from traffic, and road works.¹ The Rambla Sur lane, 4.7 kilometres long, cost close to 68 million pesos by the authorities' account.²

The size of the network is on record too: 71.3 kilometres of cycling infrastructure in 2023, under 2% of the city's street network, in a 2024 paper by Nesmachnow and Hipogrosso of the Universidad de la República, the country's main public university — a paper that is not a cost-benefit study, and that remarks that “sophisticated and complete cost–benefit analyses for cycling infrastructure are relatively limited”.²⁴

The other side of the ledger is the return. As of 3 October 2026, no cost-benefit study of Montevideo's bike lanes was known to have been published — no account that sets the 49 million or the 68 million pesos against what the lanes have given back.

The studies from elsewhere show what such an account would have to contain. A before and an after, as in Toronto and New York. Comparison streets, without which a neighbourhood's rise is credited to a lane. Sales records where they exist, since a survey records what people say they spent. A plain statement of what is counted and what is left out. And a separate line for the businesses that depend on the car, where the more robust studies point to losses.