Purpose

In the era of big data investors deal every day with a huge flow of information. Given a model populated by economic agents with limited computational capacity, the paper shows how “too much” information could cause financial markets to depart from the assumption of informational efficiency. The purpose of the paper is to show that as information increases, at some point the efficient market hypothesis ceases to be true. In general, the hypothesis cannot be maintained if the use of the maximum amount of information is not optimal for investors.

Design/methodology/approach

The authors use a model of cognitive heterogeneity to show the inadequacy of the notion of market efficiency in the modern society of big data.

Findings

Theorem 1 proves that as information grows, agents' processing capacities do not, so at some point there will be an amount of information that no one can fully use. The introduction of computer-based processing techniques can restore efficiency, however, also machines are bounded. This means that as the amount of information increases, even in the presence of non-human techniques, at some point it will no longer be possible to process further information.

Practical implications

This paper explains why investors very often prefer heuristics to complex strategies.

Originality/value

This is, to the authors’ knowledge, the first model that uses information overload to prove informational inefficiency. This paper links big data to informational efficiency, whereas Theorem 1 proves that the old notion of efficiency is not well-founded because it relies on unlimited processing capacities of economic agents.

中文翻译：

---

应用于完全有效金融市场的信息过载理论

目的

在大数据时代，投资者每天都在处理大量的信息流。给定一个由计算能力有限的经济主体组成的模型，该论文展示了“太多”信息如何导致金融市场偏离信息效率的假设。本文的目的是表明，随着信息的增加，有效市场假设在某些时候不再成立。一般而言，如果最大信息量的使用对投资者而言不是最优的，则无法维持该假设。

设计/方法/方法

作者使用认知异质性模型来说明市场效率概念在现代大数据社会中的不足之处。

发现

定理 1 证明，随着信息的增长，代理的处理能力不会增加，因此在某些时候会有大量的信息没有人可以充分利用。引入基于计算机的处理技术可以恢复效率，但机器也受到限制。这意味着随着信息量的增加，即使存在非人类技术，在某些时候将不再可能处理更多信息。

实际影响

本文解释了为什么投资者通常更喜欢启发式而不是复杂的策略。

原创性/价值

据作者所知，这是第一个使用信息过载来证明信息效率低下的模型。本文将大数据与信息效率联系起来，而定理 1 证明了旧的效率概念并不成立，因为它依赖于经济主体的无限处理能力。