When traditional markets won’t work, are costly government interventions the only option to deliver services to vulnerable populations?

Flickr.com/Taro Taylor

Getting children to school safely can be a challenging task for any parent. In Australia, parents of children with special needs receive the welcome government service of transit to appropriate schools, but in a cumbersome, time-consuming form: because the transit has to accommodate the needs of any disabled student, students who are blind or deaf ride the same buses of those who need full wheelchair access. The result is a one-size-fits-all solution that is costly to the government — and to families whose children have to spend up to four hours a day on the bus as it winds its way through its stops.

Where free markets function well, they can drive down costs while improving the quality of service, but many planners believe that in cases like this, with their complexity and small scale, an administrative solution is the best way to provide the needed service. Charles Plott, a professor of economics at the California Institute of Technology, believes that market solutions can be found for even the thorniest of “thin market” problems by combining rigorous mathematical theory and computer power with laboratory experiments, leading to an adaptable roll-out in real world conditions.

Plott and his team are using “smart markets” to give Australian parents and schools a lower-cost, higher-quality, adaptable solution to help special needs children get to their schools in comfort and safety. “We’re starting with a really, really hard case,” Plott says. “But if you can solve that case and you have happy customers, there’s lots of places for this kind of thing.”

A cascading auction

Plott’s solution begins in the laboratory, where his team has developed a series of market simulations that use the choices of incentivized human participants to test whether an efficient marketplace can be created. Optimizing the right combination of buses, vans, and other transit modes would be computationally difficult even if administrators had full knowledge of the costs of each individual transit operator — which they don’t. Instead, the price is optimized by means of a cascading auction in which individual operators submit bids and receive immediate feedback as to whether they are still in the running. In the background, the optimal combination of bidders is constantly being recalculated. The final result is a set of winning bidders that comes very near to the true optimal cost.

To translate the experiment from the laboratory to the field has required an entirely different kind of information-gathering, as stakeholders ranging from parents and administrators to transit operators and drivers’ unions have been surveyed to make sure that the Plott group’s models can work in a web of real-world contingencies. What happens if a bus doesn’t show up? What happens if a child moves out of or into the area? “All of these situational problems are challenges,” Plott says. “We’re learning how to build a robustness into these systems and to format them so that they can change to accommodate.”

The goal is that by the project’s end there will be a working system in place for the participating schools that saves money, helps parents, and serves a proof of concept for solving other problems that seem too thorny for traditional markets. Possible applications of Plott’s approach range from the allocation of public housing to moving shipping containers efficiently out of busy ports. “With proper technology,” Plott says, “I think we’re going to be able to see really major changes in the way that the public sector interacts with the private sector and in the use of competition.”