In France, outside of the Paris core, cycling is still stuck in low gear. Just 3.5% of daily commutes are by bike, even though the country has spent years expanding infrastructure and promoting active transport. A new large-scale study, “Impact of cycle lanes and public bike parking on bicycle commuting: How to achieve France’s cycling mode share target?” (Alimo et al., 2026), suggests the problem is not whether infrastructure works, but how much of it is needed, and where it is placed.
The researchers set out to answer a blunt policy question: can France realistically hit its goal of 12% cycling mode share by 2030, and what would it actually take in cycle lanes and bike parking to get there?
Their answer is both encouraging and sobering.
Infrastructure works, but not evenly
Across 6.7 million commuting records, the study finds a clear relationship between cycling infrastructure and cycling rates. Cycle lanes and public bike parking both increase the likelihood of people commuting by bike, but their impact varies significantly depending on density and location.
As the authors put it, “a one-unit increase in the rate of cycle lanes increases bicycle commuting by approximately 21% and 15% in the municipality of residence and work, respectively.”
But that headline hides a more nuanced geography.
In high-density cities, the workplace matters most. The study finds cycling is especially responsive to infrastructure near jobs, with cycle lanes at work increasing cycling probability by more than 20% in dense urban areas. In low-density areas, the effect flips: home-side infrastructure matters far more, with cycle lanes around residences increasing cycling likelihood by about 65%.
Parking helps too, but it plays a supporting role rather than a driving one. An increase in public bike parking leads to smaller gains, roughly 5% at home and 1% at workplaces across the full sample.
Or as the authors summarize, infrastructure “exerts a statistically significant and positive influence on cycling preferences, both at home and at work, and this holds across all density levels.”
What it would take to hit the 12% target
The study doesn’t stop at correlation. It models what would be needed to actually meet France’s national cycling goal.
The conclusion is stark: incremental change will not be enough.
To reach 12% cycling mode share nationally, France would need to more than double its cycle lane network or increase parking supply dramatically. In some scenarios, parking would need to increase 17-fold if used as the main lever.
Even combined, the challenge is significant.
The authors note that “reaching the national target of 12% bicycle modal share would require increasing the supply of cycle lanes and parking spaces by twofold.”
Broken down spatially, the differences become even more dramatic. High-density cities could reach the target with roughly a doubling of cycle lanes. Medium-density areas would need around five times more infrastructure. Rural and low-density regions would require increases so large they become “difficult to reach with the parking lever alone.”
In other words, France does not have one cycling problem. It has several, depending on where you are standing.
Bicycle and pedestrian bridge in Albi, France (Ney and Partners, Vincent Boutin)
Methodology
The study combines French national census data with spatial datasets on cycling infrastructure from geographic databases and OpenStreetMap, covering 6.7 million commuters across mainland France.
A model is used to estimate the probability of cycling versus other commuting modes, with a correlated random-effects (CRE) specification to account for unobserved municipal characteristics such as local cycling culture and clustering of individuals within municipalities.
Cycle lane density and public bike parking supply are measured for both home and workplace municipalities, and a range of socio-demographic controls are included. The model is estimated separately across high-, medium-, and low-density municipalities, and results are used to simulate how changes in infrastructure would affect cycling mode share under different investment scenarios.
Who cycles, and where
Beyond infrastructure, the study reinforces familiar patterns.
Men cycle more than women. Higher education strongly increases cycling likelihood. Car ownership is one of the biggest barriers, especially in lower-density areas, where having multiple cars sharply reduces the probability of cycling to work.
Distance still dominates everything. As expected, the further people live from work, the less likely they are to cycle.
But even here, infrastructure softens the edges. Better networks reduce the friction of distance, particularly in cities where trips are shorter and more direct.
The bigger message for cycling cities
The most important takeaway is not just that infrastructure matters. It is that context determines which infrastructure matters most.
In dense cities, building lanes near employment centres appears to be the strongest lever. In suburban and rural areas, home-side connectivity becomes more important. Parking helps everywhere, but it is not a substitute for connected networks.
The authors also caution that their simulation likely underestimates cycling potential. They note that the model does not fully capture “network effects,” where expanding infrastructure creates compounding benefits through safety, continuity, and social acceptance of cycling.
That matters, because cycling is not just a linear system. It behaves more like a network that strengthens as it grows.
The policy reality
France has committed billions to cycling infrastructure and plans to expand its network dramatically. But this study suggests that simply adding kilometres is not enough. The distribution, type, and density of that infrastructure will determine whether it moves the needle.
There is also a political edge to the findings. If governments focus only on high-density cities, cycling rates will rise faster there, potentially widening spatial inequality between urban and rural regions.
The authors warn of this directly, noting that “spatial equity challenges may emerge between densely populated regions and other areas.”
The bottom line
France’s 12% cycling target is not unrealistic in principle. But it is not a single lever problem either.
As the study makes clear, cycling growth depends on a layered system of lanes, parking, density, and commuting geography. Or as the authors put it, reaching national targets will require “simultaneously increasing cycle lanes and parking spaces,” rather than relying on one solution alone.
For cities and countries watching closely, the message is simple. Infrastructure works. But where you build it matters just as much as how much you build.