The Price of Stability in Matching Markets

This paper studies the inefficiency of one-to-one matching markets as measured by the price of stability. We begin by providing some theoretical upper bounds on this type of inefficiency, bounds that vary with the composition of participants’ ordinal preference lists. We then turn to simulation experiments to further describe how changes in basic characteristics of agents’ preferences can increase or decrease the efficiency of stable matchings. Our results have important implications for those who seek to improve the functioning of real-world matching markets. Though it may be difficult or even impossible to completely ascertain preferences in a real-world market, it is possible to get a sense of general levels of correlation and intercorrelation from an empirical sample. Our results can then be of help to market designers, letting them know how substantial the price of stability is likely to be.