The 10-Cent GPS

We know that technology is not a panacea, that gadgetry and software are not always the right solutions for our transport problems. But how do we know – really know -- when technology is truly the wrong way to go – when, say, using an old-fashioned compass is genuinely better than a GPS?

Thanks to blogger Sebastiao Ferreira, writing for MIT’s CoLab Radio, I have learned about an intriguing phenomenon in Lima, where entrepreneur data collectors, named dateros, stand with clipboards along frequented informal microbus routes, collecting data on headways, passenger counts, and vehicle occupancy levels. The microbus drivers pay dateros about 10-cents per instant update, and they use the information to adjust their driving speed.  For example, if there is a full bus only a minute ahead of the driver’s vehicle, the driver will slow down, hoping to collect more passengers further down the route. In informal transit systems, where drivers’ incomes are directly tied to passenger counts, paying dateros is a good investment (Photo from MIT CoLab Radio).

If you think about it, use of dateros could be more efficient than traditional schedule or GPS-based dispatch, because the headways are dynamically and continuously updated to optimize the number of passengers transported at any given time of day.  According to Jeff Warren (a DIY cartography pioneer), the dateros have been praised as the “natural database, an ‘informal bank’ of transportation optimization data.”

Does this little-known practice call into question our traditional prescription for high-tech solutions to bus dispatch?

 In trying to learn more about these dateros, I came across a fascinating study comparing the merits of this system relative to the traditional formal system of dispatch and fixed driver wages, The War of the Fare (PDF). The study, based on surveys and empirical evidence gathered in Santiago, Chile, concluded that while passengers waited about 10% longer for buses on traditionally-dispatched routes, these passengers were considerably safer – non fixed-wage routes supported by dateros, where drivers have incentives to speed and overtake other vehicles, experienced about 67% more accidents per kilometer driven.

Datero in Lima (Photo from MIT CoLab Radio)

The logical conclusion from all of this, then, would be to say that use of dateros, combined with enhanced safety enforcement, may produce optimal informal microbus transit service. But I think we can do better than that.

Let’s look at the data – this treasure trove of continuously updated headway and passenger data, whose useful life is only as long as the time the data collection page remains on the top of the datero’s clipboard.  What happens to the data after their useful life? Are the data that were worth 10-cents for an instant, truly worthless the next?  Because it would seem that these data are actually worth a lot more – more to the drivers who receive it, more to the passengers, and more to the city as a whole. But only with the risky (!) addition of a little technology.

Imagine if we were to implement this current system, but rather than paper clipboards, the dateros entered data into GPS-enabled devices. And suppose that each microbus were similarly fitted with a GPS/communications unit (as a requirement for a route license, for exampe). What would we have then?

• Passenger counts could be made publically available to the city, enabling microbus drivers considering entering a particular route to have accurate picture of what their revenue potential would be, thereby optimizing the spread of drivers across different routes. Similarly, current drivers would be able to compare revenues against other drivers, providing positive reinforcement that with dispatch, everyone wins;
• If linked with a crowd-source mapping initiative, like Todos Somos Dateros in Lima, then we would have further information about specific safety and security issues, in real time, by vehicle – again, improving the city’s ability to enforce and improve service;
• Driver dispatching based on passenger flows would be instantaneous (no need for dangerous jump-on’s and jump-offs by the dateros);
• Formalized stations could be created based on an unprecedented data sample and adherence to use of these stations could be enforced (through vehicle GPS-tracking combined with the passenger count data);
• If the city hoped to transition in the future to a formalized transport system, developing optimal routing and scheduling would be straightforward;
• Safety violations (erratically-driven vehicles, speeding, etc.) would be instantly logged, and drivers could be fined from a centralized system, thereby improving safety;
• General traffic flow conditions could also be recorded through the GPS-enabled microbuses, providing transport planners with accurate pictures of congestion;
• Etc. etc. etc. These are the ideas I have off of the top of my heads. Would be interested in yours…

With this combined manual data collection + open technology system, we avoid the Prisoner’s Dilemma that has plagued efforts to institute microbus dispatch in other cities, such as Manila, where asymmetric information and introduction of a manual dispatch system resulted in  corruption. Along the busy Manila EDSA corridor, street-side dispatchers guide microbus drivers on a time-based schedule, ensuring relatively predictable headways for users and fair distribution of revenues for the drivers.  As in the classicial prisoner's dilemma, drivers have the option to comply or to cheat (i.e., bribe the dispatcher to ensure greater passenger capture for themselves), and drivers cannot readily know what choice other drivers have made.  If all drivers comply, everyone wins, if all drivers cheat, everyone loses, and if only some drivers cheat, we find gross imbalances in service and revenue -- as is the case today along EDSA.

So with open information and a little technology, we can avoid this dilemma, while at the same time improve service and, potentially, revenues. But then, by supplementing the existing system, maybe we cause a new dilemma –what incentives could we put in place to ensure the dateros continue to collect the best data possible? What payment system would be most effective?

Also, given the low cost of the dateros’ function, it would seem that transport planners would need to do a little soul-searching about whether our new-fangled technical solutions to age-old data collection and dispatch problems are truly warranted. A datero report is only 10-cents. A real GPS system costs…a lot more. How much more would need to be charged per ticket for this safer, more secure, better level of service? Or more to the point, how much more would passengers be willing to pay?

Is this an instance where the technological solution is better?

The joyful thing is that we could, conceivably, answer this question. With a little grant funding and a set of survey questionnaires, we can ask passengers how much more they would be willing to pay for better service, drivers the extent to which they believe they could benefit from the system, dateros about effective incentives.  Keeping my eyes open for the next trust fund application cycle – and some interested summer interns….

Special thanks to Nick Grossman at Open Plans for leading me to Sebastiao’s post.

All images from Sebastiao Ferreira’s blog at MIT’s CoLab Radio.

Full Citation for The War of the Fare: Ryan M. Johnson, David H. Reiley, and Juan Carlos Muñoz. February 2006. “The War of the Fare.” http://economics.uchicago.edu/pdf/Appl_Dynmcs_Reiley_5-18-06.pdf