Information Casacades in Markets

这是女性研究机会/研究计划基金。信息级联是一种有趣的现象，受到了经济理论家的广泛关注。当个人拥有私人信息并按顺序做出公共决策时，可能会形成级联。如果早期的决定是一致的，那么其他人效仿它可能是最优的，而不管他们的私人信息是什么。关于级联效应的轶事例子包括时尚、金融市场趋势以及民意调查对投票行为的影响。在大多数自然发生的情况下，很难排除顺从的其他解释。来自个人决策实验的数据（Anderson， 1994年和Anderson and Holt， 1996年）表明，在一个受控的实验室环境中，其他追随的动机被最小化，级联形成。本研究在更丰富的市场环境中引入了信息级联的可能性。以往的实验研究集中在信息传播揭示资产真实价值的情况下，从而使交易者受益。在拟议的实验中，交易员拥有私人信号，这些信号提供了有关资产价值的不完整信息。交易按顺序进行，并公开宣布（例如，价格出现在自动收报机上）。在某些情况下，这些公开交易会泄露私人信息。如果早期交易者被误导，可能会形成一个反向的连锁反应，其他交易者跟随错误的决定，即使他们的私人信息会引导他们做出相反的决定。该模型的参数化版本的试验会话显示，在许多可能的情况下，会形成级联和反向级联。以下初步研究将在本研究计划资助期内进行：建立理论模型、确定有趣的实验治疗方法、进行试点实验，并根据试点实验的结果准备正式的资助建议。

Lisa Anderson This is a Research Opportunities for Women/Research Planning Grant. Information cascades are interesting phenomena that have received considerable attention from economic theorists. Cascades may form when individuals have private information and make public decisions in sequence. If early decisions match, it may be optimal for others to follow, regardless of their private information. Anecdotal examples of cascades include fads, trends in financial markets, and the influence of polls on voting behavior. It is difficult to rule out other explanations for conformity in most naturally occurring situations. Data from individual decision making experiments (Anderson, 1994 and Anderson and Holt, 1996) show that cascades form in a controlled laboratory environment where other motives to follow are minimized. This research research introduces the possibility of information cascades into a richer market setting. Previous experimental studies focus on situations where information dissemination reveals the true value of the asset, and hence, benefits traders. In the proposed experiments, traders have private signals that provide incomplete information about the value of an asset. Trades are made in sequence and are publicly announced (e.g., prices come across a ticker tape). In some cases, these public trades reveal private information. If early traders are misinformed, a reverse cascade may form, where other traders follow with incorrect decisions even though their private information would lead them to do otherwise. A pilot session of a parameterized version of this model reveals that cascades and reverse cascades form in many cases where they are possible. The following preliminary research will be conducted during the period covered by this Research Planning Grant: developing a theoretical model, determining interesting experimental treatments, conducting pilot experiments, and preparing a formal grant proposal based on the findings of the pilot experiments.