Emissions Permits Auctions: An Agent Based Model Analysis

Abstract

In this chapter the authors demonstrate with three relevant issues that Agent Based Modeling (ABM) is very useful to design emissions permits auctions and to forecast emission permits prices. They argue that ABM offers a more efficient approach to auction design than the usual mechanistic models. The authors set up the essential components of any market institution far beyond supply and demand. They build an ABM for the emissions permits auction of the Environment Protection Agency (EPA), and demonstrate why the EPA failed. In the second experiment they show that in a competitive and efficient auction, the Continuous Double Auction, there is room for traders learning and strategic behavior, thus clearing the perfect market paradox. In the third experiment they build an ABM of the Spanish electricity market to get CO2 emissions prices forecasts that are more accurate than those obtained with econometric or mechanistic models.