As a cyclist, I believe that bike lanes do induce demand, but there are some important differences. A single road can attract travels which begin or end far aw

As a cyclist, I believe that bike lanes do induce demand, but there are some important differences. A single road can attract travels which begin or end far away from the road. So its effect is felt on a larger territory and further away. A road needs only few exits to be connected to other roads. Bike lanes are more like capillary network than arteries. A single bike lane is almost useless, because it can serve only travels which begin and end in a narrow strip around the lane (maybe a block away). Two bike lanes which are not connected will be used much less than two connected lanes. Even a single discontinuity of the network can make bike travel unfeasible. So rather than talking about lanes, we should talk about lane networks and their density. And up until certain density the network won't be used much. Only after reaching that threshold we can talk about effects it may have on city development. Similarly, a car road will not generate induced demand nor attract much traffic if it is missing a bridge across the river. Let say a bike lane attracts traffic if the starting point and the destination are within w/2 from the lane. Let's assume uniform distribution of potential starting and destination points. Then in a city area A, a bike lane of length L will attract only Lw/A potential commutes (within a strip around the lane). There are two important observations to be made: 1) the induced demand is proportional to the length of the network L, 2) parameter w is probably small, likely ~ 200m, so 1 km of a bike lane will have a smaller and more local effect than 1 km of the road. To cover 1 km² of urban area, the well connected network would have to be a grid with the total density of bike lanes ~ 10 km / km².