Predictive Agent Class Collusion And Its Regulatory Path Analysis

In recent years,the pricing algorithm technology has become more and more complex,and technological progress has achieved the transformation from mechanical pricing to artificial intelligence pricing algorithm.The artificial intelligence algorithm can process a large amount of data,and interact,learn and make decisions with unprecedented speed and complexity.The development of algorithmic pricing software has attracted people’s attention to its potential impact on corporate behavior and competition.In particular,the potential of using algorithms as a means of promoting collusion,whether implicit or explicit,has been a hot topic of discussion in recent years.For the collusion caused by algorithm,the traditional regulation is mainly aimed at explicit collusion,and for implicit collusion,there is no corresponding anti-monopoly tool.How to build a long-term regulatory mechanism to deal with implicit collusion is a major problem in anti-monopoly.At present,academic circles still have different views on the impact of pricing algorithms on competition,regulatory path,etc.Therefore,this thesis focuses on the formation,operation mechanism,characteristics,manifestations and corresponding policy recommendations of algorithms to promote collusion.The existing literature on pricing algorithm collusion focuses on the impact of pricing algorithm collusion on competition as a whole,and less on the specific types of pricing algorithm collusion;On the other hand,it focuses on theoretical analysis and does not use models to verify to increase persuasiveness.In view of the above limitations,this thesis selects the infinitely repeated game model in the research method,and proves the inevitability and stability of predicting the existence of agent collusion under the premise of high market transparency through model reasoning.In terms of research content,it focuses on the specific type of pricing algorithmic collusion-forecasting agent-type collusion,focuses on the analysis of the formation mechanism,operation principle,manifestation,characteristics,and the necessity of regulation of forecasting agent-type collusion,and on this basis summarizes the regulatory dilemma of forecasting agent-type collusion,and puts forward corresponding policy recommendations for the actual law enforcement dilemma.After theoretical analysis and game model analysis,this thesis puts forward the following conclusions:(1)The use of pricing algorithm will increase the possibility of collusion between enterprises;(2)It is predicted that agency collusion will have higher concealment and stability,which greatly increases the difficulty of anti-monopoly law enforcement.Based on this,in view of the characteristics of high stability,strong concealment and intelligence of the prediction agent collusion,combined with the impact of collusion on consumer welfare,the damage to competition order and other hazards in reality,and its regulatory dilemma,this thesis puts forward targeted policy recommendations,including:(1)the pre-regulation of the prediction agent collusion: innovate the pre-regulation law enforcement tools,strengthen the compliance review of the pricing algorithm,and clarify the subject of legal responsibility;(2)Predict the process supervision of agency collusion;(3)Anticipate the post-event antitrust of agency collusion.It is hoped that it can provide reference for the anti-monopoly regulation of prediction agent collusion in the future.