Slate has an article that asks an intriguing question: Who Really Benefits From “Big Data”? (Dec 27). That’s clearly something of a loaded question. Big Data is currently everyone’s favorite answer for everything. The ability to leverage vast amounts of data for new insights and improved decisions holds a lot of promise. There are also many success stories of firms creating new markets or improving profitability or providing great value to customers to back up claims and bolster expectations. Big Data has remade baseball with an emphasis on new statistical measures and has allowed Netflix to suggest the perfect next movie to watch.
Those examples sound great. Of course, consumers may be less enthusiastic about one of the longer standing examples of Big Data, airline revenue management systems. While these have been around for a couple of decades,they bear the hallmarks of Big Data applications. They are built on careful data analysis to forecast how systems will evolve and seek to replace intuition with frequent, reasoned decisions. These decisions may not necessarily be optimal but they clearly balance costs and benefits and can be improved over time. I’m not sure that customers love revenue management systems the way they love Netflix recommendations. Although revenue management systems are just as responsible for some sweet deals as they are over for extravagantly priced tickets, people tend to focus on the latter. Consequently, if you ask who benefits from Big Data and lead with revenue management systems as an example, I would venture that many customers would be leery of embracing Big Data.
So what example does the Slate article go with in thinking about Big Data? Lexus Lanes on the DC Beltway!
Advances in real-time data acquisition, processing, and display technologies means that it is possible to design a toll road that can continually change prices to control how many cars are on the road and how fast they are going. These “hot lanes“ have just been opened along a part of the Washington, D.C., Beltway, the 10-lane, traffic-infested artery that to normal humans is a metaphorical boundary between the real, outside-the-Beltway world and the weird, political one on the inside. (For those of us who live around Washington and must drive on it, however, the Beltway is very concrete indeed, a daily flirtation with delay and frustration, homicidal instincts, and death itself.)
At a cost of $2 billion, a private sector partnership (which gets to keep the tolls) has built a 14-mile-long, four-laned section of highway, parallel to the main lanes of the toll-free Beltway, and has guaranteed to the state of Virginia that it will always keep traffic moving at no less that 45 mph along its length. They do this by continuously monitoring the number of cars (which must be equipped with EZ-Pass transponders) and their speed, and by raising toll prices as necessary to keep the number of cars on the road at a level that will allow the speed to stay at or above the guaranteed minimum. The dynamic toll prices are displayed on huge signs near the entrances to the smart-highway lanes, so drivers get to decide at the last minute whether they want to spend the money to go faster or not. As the traffic on the toll-free Beltway lanes gets worse, some drivers will be willing to spend more to go faster. The worse the traffic is, the more they’ll have to spend. (In the early days of this new technology, numerous accidents were caused by drivers trying to decide how much they were willing to pay, but no doubt this initial problem will sort itself out as people get used to driving-while-economically-rational.
We have written about such systems in the past and have raised similar points to the article. Most importantly, it is not clear that the value of one’s time necessarily translates into an ability to pay. On a given day, some driver might have a real need to move fast but that clearly does not translate into a greater ability to pay for the privilege. Once we throw in a budget constraint, you don’t necessarily get the most efficient allocation of capacity in the sense that those with the highest cost of waiting are not necessarily those in the express lane.
Having said that, it’s worth pointing out something that the article ignores. These new lanes in DC have at least expanded capacity. That is, the lanes dedicated to express traffic wouldn’t be there without the special tolls. So it is not apriori clear how this plays out for the traveling public at large. On the one hand, users of the standard Beltway lanes are certainly worse off than they would be if all users shared all available lanes. On the other, they are better off than they would be if all users were forced to use just the standard lanes.
A final point. Pricing access to these express lanes is not as straightforward as you might think. The tricky part is that the current price of the express lanes affects not only how much the operator makes right now but also how much the operator makes in the future since current pricing impacts how congested each option will be in the future. Some colleagues at Columbia have done some work on this problem. One of their findings is that it is optimal to raise prices earlier than would immediately seem appropriate. Think about early in the morning just as rush is about to start. If all lanes are moving well, it would seem that the only way the express lanes could get business would be to price low. Not pricing low seems be leaving money on the table since the lanes are otherwise underused.
That’s not what’s optimal, however. Optimal pricing calls for charging a pretty stiff premium even if standard lanes are moving well. The key is that the operator wants the standard lanes to be a terrible option once rush hour is properly underway. That means cramming as many cars onto the standard lanes early in rush hour so when there are a lot of potential express lane customers, there is a compelling reason to take them.