Friday, March 17, 2006
The Power of Incentives -- Chilean Bus Driver Edition
In a column on Slate yesterday, Chicago economist Austan Goolsbee asks why, when traffic is bad, do bus drivers sit in traffic on Lake Shore Drive (where there are no stops) while cab drivers get off and take shortcuts?
Unsurprisingly (this is an economics column after all) incentives provide the answer. Chicago's bus drivers are paid per hour. While their passengers lose when the bus sits in traffic needlessly, the drivers don't. In fact, thinking about how to get around the traffic only imposes effort costs on the drivers. Thus, the drivers face a disincentive to improve their riders welfare. In Chile, though, bus drivers are not paid per hour, but rather per passenger. As such, they run on time.
Goolsbee describes an interesting market with some interesting features. E.g., the drivers pay people to wait at key stops in order to provide information about their location relative to other busses along the route. And Goolsbee correctly points out that this scheme increases driver productivity and that government should consider such approaches because they can generate the same amount of service for less money.
I have a couple of concerns, though, about whether or not this scheme actually improves social welfare that Goolsbee doesn't fully address. First, paying drivers in this manner also incentivizes them to drive aggressively and recklessly (see cab drivers -- who are already paid per passenger). Goolsbee notes that this problem exists, but suggests that passengers accept the increased tension and nausea because the per passenger bus companies do better then per hour bus companies. However, Gooslbee fails to mention the externalities reckless bus drivers impose on other drivers and pedestrians. I do not know how big these costs are, but before implementing a policy one should try and figure this out.
Second, I am curious how the companies or drivers solve a coordination problem. Drivers want to run far enough behind the previous driver on their route to allow more passengers to build up (hence the people who wait at stations to tell them how close they are to the previous bus). However, passengers (and probably drivers) want the line to run on time. Regular riders want to be able to roughly time their arrival at pick up points to coincide with the arrival of the bus (particularly when the weather is bad). Yet, if a driver receives word that the driver ahead is behind schedule, he now has an incentive to delay his route in order to not follow closely behind. While all drivers (and the bus company) benefit from having a well coordinated schedule, each individually has an incentive to cheat the system in order to maximize their own ridership. Thus, the company would need to design a system to offset this problem.
Finally, most bus routes that I have observed don't have a great deal of flexibility. That is, it seems like there are relatively few routes which have sufficiently long hauls to allow a driver choice of route while still making all their stops. E.g., the number 1 bus runs up and down Mass. Ave. stopping every couple of blocks. There doesn't seem to be much of a margin to increase these drivers productivity via pay incentives. In fact, such a system would likely only introduce volatility into workers pay, and most people don't like such risk.
Unsurprisingly (this is an economics column after all) incentives provide the answer. Chicago's bus drivers are paid per hour. While their passengers lose when the bus sits in traffic needlessly, the drivers don't. In fact, thinking about how to get around the traffic only imposes effort costs on the drivers. Thus, the drivers face a disincentive to improve their riders welfare. In Chile, though, bus drivers are not paid per hour, but rather per passenger. As such, they run on time.
Goolsbee describes an interesting market with some interesting features. E.g., the drivers pay people to wait at key stops in order to provide information about their location relative to other busses along the route. And Goolsbee correctly points out that this scheme increases driver productivity and that government should consider such approaches because they can generate the same amount of service for less money.
I have a couple of concerns, though, about whether or not this scheme actually improves social welfare that Goolsbee doesn't fully address. First, paying drivers in this manner also incentivizes them to drive aggressively and recklessly (see cab drivers -- who are already paid per passenger). Goolsbee notes that this problem exists, but suggests that passengers accept the increased tension and nausea because the per passenger bus companies do better then per hour bus companies. However, Gooslbee fails to mention the externalities reckless bus drivers impose on other drivers and pedestrians. I do not know how big these costs are, but before implementing a policy one should try and figure this out.
Second, I am curious how the companies or drivers solve a coordination problem. Drivers want to run far enough behind the previous driver on their route to allow more passengers to build up (hence the people who wait at stations to tell them how close they are to the previous bus). However, passengers (and probably drivers) want the line to run on time. Regular riders want to be able to roughly time their arrival at pick up points to coincide with the arrival of the bus (particularly when the weather is bad). Yet, if a driver receives word that the driver ahead is behind schedule, he now has an incentive to delay his route in order to not follow closely behind. While all drivers (and the bus company) benefit from having a well coordinated schedule, each individually has an incentive to cheat the system in order to maximize their own ridership. Thus, the company would need to design a system to offset this problem.
Finally, most bus routes that I have observed don't have a great deal of flexibility. That is, it seems like there are relatively few routes which have sufficiently long hauls to allow a driver choice of route while still making all their stops. E.g., the number 1 bus runs up and down Mass. Ave. stopping every couple of blocks. There doesn't seem to be much of a margin to increase these drivers productivity via pay incentives. In fact, such a system would likely only introduce volatility into workers pay, and most people don't like such risk.
// posted by BW @ 8:28 AM 
Subscribe to Posts [Atom]