错误地勾结算法复杂性驱动力是否具有竞争力

IMF | Statistics1Collusion by mistake: does algorithmic sophistication drive supra-competitive profits? NOVEMBER 21, 2024Ibrahim AbadaGrenoble Ecole de ManagementIn collaboration with Xavier Lambin and Nikolay Tchakarov (ESSEC Business School)IMF | Statistics2A DisclaimerThe opinions expressed in this presentation are those of the authors alone and might not represent the views of GEM or ESSEC. IMF | Statistics3Context▪The literature consistently reports that simple reinforcement learning algorithms systematically reach seemingly collusive outcomes.▪The drivers of cooperation are being investigated: sophisticated punishment strategies to sustain the cartel (Calvanoet al. [2002b]), numerical biases (cooperation bias Banchioand Mantegazza[2023]), correlated learning (Lambin [2024]), etc. ▪Often simple Q-learning algorithms are tested with an implicit asusmption: “The enhanced sophistication of learning algorithms makes it more likely that AI systems will discover profit-enhancing collusi