Free transit won't keep commuters coming back

Just a quick note. Eric Jaffe has a review of an MIT study of transit mode shift. This study has a lot of interesting implications, but claims of the value of free transit are overstated.

The study piggybacked on a MIT initiative in September 2008, where all employees with full-time parking permits were offered a free one-week transit pass. A number of these employees agreed to participate in the study, where they committed to commuting by transit for at least three days that week, and fill out a survey. Thirty percent of the survey participants gave up their parking permits immediately, and 25% were still on transit six months later.

Jaffe touts the free aspect of the trial as significant, but I think he - and the study authors - are overstating it. The key is that the transit was only free for one week; after that, the participants who shifted to transit had to pay full price.

Remember, these are full-time MIT employees; I don't think they're exactly poor. In the surveys they indicated that relative cost was a factor in their mode choice, but transit doesn't have to be free to compete on cost grounds. Most of the transit passes mentioned by the authors cost more than $65 per month, but probably less than $65 plus gas, car maintenance and insurance.

There are important take-aways to this study. I'm sure the free trial helped convince drivers to try transit; it's hard to justify even a week of transit if you're already paying $65 per month for a parking permit. And once they tried it, many of them realized that it wasn't all creeps and weirdos. Equally important, there was a value proposition there. For some participants, transit provided a better overall value for the price than driving. For other participants it didn't provide better value, and they went back to driving.

Transit managers that believe they do offer a better value than drivers should try free trials. Rather than offering them to everyone and overloading the system for a week, though, it would probably be better to simply mail a random chunk of the population free one-week transit passes. Whatever happens, it's important to make a clear distinction between free trials and free transit.