Abstract

Jean-Philippe Bouchaud's suggestion to stress test economic and financial instruments "in vitro" before making them available to the general public recalls recent developments in the field of experimental economics aimed at supporting the design of financial instruments and economic institutions at large. The main advantage of the experimental laboratory, besides the possibility to control all the variables that are relevant to the situation at hand (being it a market, a twoperson bargaining game, or any other economic setting worth of being analyzed through experiments) is the opportunity to test the effects of institutional changes in small-scale "safe" conditions before introducing those changes in the real world, thus helping avoid potentially catastrophic mistakes. The use of experiments for design purposes has a long history in experimental economics, and it has recently regained popularity due to a combination of factors: firstly, many contemporary markets and other institutions are so complex that the definition of optimal mechanisms cannot resort to theoretical tools alone. Secondly, the experimental evidence has convincingly demonstrated that even apparently irrelevant details of the microstructure of interactions can have an enormous impact on aggregate efficiency. The article briefly reviews some evidence on experimental design economics as applied to both markets and organizations. In the first domain, the applications discussed include electricity markets, online auctions, financial markets, and the design of call options. In the organizational domain, we discuss results from experiments on the effect of different compensation schemes and career plans, and on the institutional and behavioural features that may enhance coordination in "weak link" types of tasks. We conclude by advocating a more extensive use of the experimental method in economics to inform the engineering of economic institutions, thus helping to avoid disastrous consequences of wrong designs in the future.