Accessing Bart

Parking spaces are expensive. The average cost for a single parking space is approximately $2,500 per year. This value increases with the high land costs in San Francisco as well as for downtown areas. This number increases as the value of land becomes more expensive. Each BART station with parking (79% of the total 42 stations analyzed) has an average of nearly 1500 spaces — this is $3.75 million dollars per year per station for parking costs!

Unlike many commuter park-and-ride lots in the Bay Area, BART station lots charge for parking. Daily fees range from $3.00 to $6.00, depending on the demand for the station. For many BART parking lots, a long wait list to purchase a reserved monthly permit exists—suggesting that prices do not reflect the market demand.

Providing parking at BART station lots has both benefits and drawbacks. Park and Ride lots may discourage users from driving to work, reducing peak-period congestion during the commute period, decreasing Vehicle Miles Traveled (VMT) and increasing transit ridership. However, without parking, many users would otherwise access stations with alternative means of transportation — such as bicycling, taking transit, or carpooling. This would reduce harmful vehicle emissions (the highest fuel burn and levels of pollutants emitted from vehicle are from warming up the engine, or “cold starts”), save public costs for BART to provide parking, and encourage other land uses nearby stations.


When the number of auto spaces is compared to the drive-alone, or SOV mode share, there is a strong correlation between the two variables (R-squared=0.66, p<.001). Just the question of whether or not there is parking at a BART station is an even stronger predictor of non-SOV mode share (R-square = 0.76, p<.001). Combining these two factors, we get the strongest prediction, with an R-square of 0.84 (statistically significant at a 99% confidence level). Click on the points in the plot below to view stations.

The map at the top of the page allows you to view the stations where there are imbalances between the percentage of individuals accessing the station with single-occupancy vehicles and the amount of parking. For example — West Oakland has a large percentage of individuals accessing with vehicles, yet only 451 spaces. This shows that users may be parking elsewhere. The map also shows where there may be an oversupply of parking. El Cerrito del Norte, for example, has a low drive-alone mode share, but a large number of parking spaces (2050).

Policy Implications

These results show that there is a strong correlation between parking and SOV mode share. However, what may be more interesting than the strength of the correlation are the outlier stations. Mousing over the figure above will show the stations that have a high number of parking spaces, yet low SOV mode shares (for example, El Cerrito del Norte and Dublin/Pleasanton stations). This suggests that these stations may oversupply parking, wasting money and land resources that could be allocated elsewhere.