As the recipients of the 2012 science Nobel prizes gather in Stockholm to celebrate and be celebrated, News & Views shares some expert opinions on the achievements honoured.
What is the best way to match up entities that have different preferences for one another, if price cannot be used to determine the allocation? The 2012 Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel went to Lloyd S. Shapley and Alvin E. Roth for the theory and practical demonstration that such processes are optimized by achieving stable matches (see Fig. 1).
by Yan Chen
In a seminal paper, Gale and Shapley1 outline an algorithm that always produces stable matching in situations in which legal or ethical obligations preclude the use of price to determine the allocations. Extensions of this algorithm have been applied to several important fields, including the labour market for doctors2, school admissions3 and kidney exchange4, to improve the stability and reduce gaming of the matching systems in use.
In such system redesigns, controlled laboratory experiments are frequently used to compare the performance of the systems at a level of detail that cannot be obtained from field data. For example, Kagel and Roth's experiments5 on regional medical markets in Britain allowed a comparison of the performance of two algorithms for assigning doctors to hospitals, and this demonstrated that the stability of the algorithm contributes to the functioning of such markets. Laboratory experiments can also generate the first data on the performance of a theoretically superior mechanism for which there are no field data. For example, the first school-choice experiment6 helped to persuade the public-school authorities in Boston, Massachusetts, to switch in 2005 from the Boston mechanism then in use to the Gale–Shapley mechanism.
These examples demonstrate how laboratory experiments can serve as a wind tunnel for evaluating new institutions. Matching and auction theories thus provide new areas for experimental research and are canonical examples of scenarios in which theory, laboratory experiments and real-world implementations form a healthy feedback loop. For these reasons, more economists are now conducting laboratory experiments to evaluate policies and institutions.
An engineering approach
by Jacob Goeree
Roth's early work illuminated how the Gale–Shapley algorithm1 led to the successful matching of newly qualified doctors to hospitals, where decentralized systems had failed. New challenges arose, however, when a growing number of couples graduated from medical school and started to contact hospitals directly, causing instabilities. So Roth and Peranson designed an algorithm that could accommodate joint applications by doctors. Many entry-level labour markets now use this algorithm, which produces stable outcomes even when the applicants are couples.
In the process, Roth came to realize that many practical problems could not be solved by theory alone, and that computer simulations and laboratory experiments were invaluable tools for comparing alternative matching mechanisms. This 'engineering' approach inspired an entirely new field, referred to as market design, which draws on insights from game theory, experimental economics and computer science to improve the functioning of economic and social institutions.
The tools of market design are now being applied to a host of settings, including auctions to privatize public assets7 and cap-and-trade programmes to reduce greenhouse-gas emissions8. In an interview about the Nobel award, a reporter asked Roth's opinion on the European debt crisis, and he modestly replied that he is not “that kind of economist”. And yet his engineering approach to market design could help to create stable financial markets that avoid excessive risk-taking and taxpayer bailouts of banks or countries, such as those that are currently plaguing economies around the world.
Gale, D. & Shapley, L. S. Am. Math. Mon. 69, 9–15 (1962).
Roth, A. E. J. Polit. Econ. 92, 991–1016 (1984).
Abdulkadiroğlu, A. & Sönmez, T. Am. Econ. Rev. 93, 729–747 (2003).
Roth, A. E., Sönmez, T. & Ünver, M. U. Q. J. Econ. 119, 457–488 (2004).
Kagel, J. H. & Roth, A. E. Q. J. Econ. 115, 201–235 (2000).
Chen, Y. & Sönmez, T. J. Econ. Theory 127, 202–231 (2006).
Goeree, J. K. & Holt, C. A. Games Econ. Behav. 70, 146–169 (2010).
Ishikida, T., Ledyard, J., Olson, M. & Porter, D. Res. Exp. Econ. 8, 185–220 (2001).