不确定条件下的个人选择——行为经济学带来的挑战

发布时间：2018-05-17 23:19

本文选题：不确定条件 + 预期效用理论　；参考：《吉林大学》2004年硕士论文

【摘要】：现今在经济研究中，探讨人们是怎样和为什么作出自己的选择已经成为一个非常迷人的领域。而这些研究则围绕着不确定条件下人们是如何针对环境和条件作出自己较为准确的选择。最广为接受的不确定条件下的个人选择理论就是由冯·诺伊曼—摩根施坦因（1953）提出的预期效用理论。然而有关这个理论的完善性的问题最近经常被提出。这些问题的提出对于真正的解释不确定条件下人们究竟是如何作出选择的是十分有推动作用的。“阿莱斯悖论”对预期效用理论形成了最先的挑战，此后建立在偏好诱导基础上的偏好逆转问题的实验发现则对其形成了真正的打击。于是，卡尼曼和特韦尔斯基给出了有关人类在不确定条件下如何作出选择的新的解释，即行为经济学的解释。预期理论应运而生，它成为了预期效用理论的一种有效的替代理论。他们发现预期效用理论不能全面的描述人们在风险条件下做出的抉择，会出现决策者选择不能够被预期的情形。例如，他们指出结构会改变个人的决定还有为什么个体在有些情形下选择风险偏好的而在另外一些情形选择风险厌恶。本文就是站在预期理论对预期效用理论替代发展的基础上，更加准确的向人们呈现出人们究竟是如何在不确定条件下作出自己的选择的。第一章以预期效用理论的产生为开篇，说明预期效用理论的假设是对早期风险承受理论的替代，通过圣比德堡悖论的提出和解决使其初步产生，继而较为详细的描述了预期效用理论模型。第二章则对预期效用理论提出了异议，并分别详细给出了阿莱斯悖论”（Allais paradox）和偏好逆转现象。“阿莱斯悖论”实质上是要揭示，许多建立在独立性假设上的期望效用，尤其是建立在追求期望效用最大化基础上的模型都忽略了人的心理因素对概率分布的影响，因而不切合实际。随后介绍了偏好逆转现象，其对效用理论形成了真正的冲击，是对经 WP=40 济理性公理之一可传递性的致命挑战。第三章针对偏好逆转问题指出建立在公理性假设基础上的预期效用理论的不足之处，其对静态公理化体系的空间和时间一致性假设提出了质疑。以及给出卡尼曼和特韦尔斯基对它提出的质疑。认为人们作出决策是很难作到完全理性的，效用理论的很多假设是站不住脚的。最后指出效用最大化模型的缺陷。主要有三点：（1）回顾效用(recollected utility)和体验效用的不一致。（2）决策结果的权重不能正确分配。（3）决策效用(decision utility)与体验效用(experienced utility)的不一致。第四章详细叙述预期理论，具体介绍人类决策的非理性和风险偏好。如：第一、确定性效应(certainty effect)，在特定情况下，人的效用函数低估一些只具有可能性的结果，而相对高估确定性的结果。第二、反射效应(reflection effect)，根据反射效应原理，人们对盈利前景的选择和对亏损前景的选择是等价的，在亏损前景间的选择就好比是在收益前景间选择的一个镜面反映(a mirror image)。第三、分离效应(isolation effect)，人在分析评估不同的待选择前景时，往往会暂时剔除各种前景中的相同因子，而通常情况下剔除方法并不是唯一的，处理问题的方法的多样性也会导致人的偏好与选择不一致，这就是所谓的偏好分离效应。以及人们进行决策时常常出现的心理状态：第一、锚定效应(anchoring effect)，人们在进行判断时常常过分看重那些显著的、难忘的证据，甚至从中产生歪曲的认识。第二、后悔理论，卡纳曼等心理学家认为，当人们犯错误时，哪怕是很小的错误，也会有后悔之痛，并会严厉自责，而不是从更远的背景中去看待这种错误。第三、过分自信(over confidence)行为经济学在研究过程中还发现一种现象，那就是人们对自己的判断都过分自信。最后给出预期理论模型，并详细指出它与预期效用理论模型的不同之处。Kahneman和Tversky并没有止步于批判，他们在其《预期理论:对风险决策的分析》(1979)一文中，，提出了另一种分析框架。预期效用理 WP=41 论是公理性的，而他们的预期理论是描述性的，是以演绎的方式从经验观察中得到的，而不是通过归纳的方式从一组逻辑上有吸引力的公理中得到的。模型的详细区别在于：首先，在预期理论中，决策者并不关心财富本身的最终值，而关心财富相对于某一参考值的变化Δw。这一参考值常常是决策者当期的财富水平，就是说货币损益根据现状定义。预期理论区别于预期效用理论的第二点与价值函数v有关。除了定义根据财富变化以外，这一函数是Ｓ型的，对收益而言它是凹的，对损失而言它是凸的，对两个方向的变化呈现敏感性递减趋势。而且，它在原点上发生弯曲。与在原点右方收益很小的情况相比，在原点左方，当损失很小时，函数更加陡峭。相反，在期望效用理论中，函数ｕ通常处处平滑并且为凹函数。第三，决策权重函数π是客观概率p和q的函数。该函数单调递增，在0和1处不连续。因此，对于小概率，它始终赋予过大的权重；而对于大概率，则始终赋予过小的权重。第五章对预期理论进行客观的评价，并指出其不足之处，以及它与预期效用理论的替代关系和角色关系。最后指出其对我国的现实借鉴意义。尽管预期理论的基本理论框架的构建有待完善，但其以更加逼近真实世界的人类行为，?
[Abstract]:Nowadays, in economic research, it is a fascinating field to explore how and why people make their own choices. These studies focus on how people make their own more accurate choices about the environment and conditions under uncertainty. The most widely accepted theory of individual choice under uncertainty is the theory of personal choice. The expected utility theory proposed by von Neumann Morgan Stein (1953).
However, questions about the perfection of this theory have recently been put forward. These questions are very helpful to the real explanation of how people choose to make a choice. "Alice paradox" has formed the first challenge to the expected utility theory, and has been built on the basis of preference induction. The experimental discovery of preference reversal has a real impact on it.
As a result, Kahneman and ravinsky gave a new explanation of how human choices were made under uncertain conditions, the explanation of behavioral economics. The expected theory came into being. It became an effective alternative theory of expected utility theory. They found that the expected utility theory could not fully describe people in the risk conditions. Decisions make decision makers to choose situations that cannot be expected. For example, they point out that structure changes individual decisions and why individuals choose risk preference in some cases and choose risk aversion in other situations.
On the basis of the alternative development of expected theory of expected utility theory, this article is more accurate to show people how people make their own choices under the uncertain conditions.
The first chapter begins with the emergence of the expected utility theory, which shows that the hypothesis of expected utility theory is an alternative to the early risk tolerance theory, which is initially produced by the proposal and solution of the St. Petersburg paradox, and then a more detailed description of the expected utility model.
The second chapter gives an objection to the expected utility theory, and gives the ales paradox (Allais paradox) and the phenomenon of preference reversal in detail. The "ales paradox" is essentially to reveal the expected utility built on the assumption of independence, especially the model established on the basis of maximizing the expectation utility. The influence of psychological factors on probability distribution is therefore not practical. Then the phenomenon of preference reversal is introduced, which has a real impact on utility theory.
WP=40
One of the axioms of economic reason is the fatal challenge of transmissibility.
The third chapter points out the deficiency of the expected utility theory based on the axiom hypothesis, and raises questions about the space and time consistency hypothesis of the static axiomatic system, and gives a challenge to it. It is difficult for people to make a decision to make a decision. Many hypotheses of utility theory are untenable. Finally, the defects of the utility maximization model are pointed out. There are three main points: (1) recollected utility and the inconsistency of experience utility. (2) the weight of decision results can not be correctly assigned. (3) the difference between decision utility (decision utility) and experience utility (experienced utility) Cause.
The fourth chapter describes the expectation theory in detail, specifically introducing the irrational and risk preference of human decision-making. For example, the first, certainty effect, under certain circumstances, the human utility function underestimates the results that only have the possibility, and the relative overestimation of the certainty of the fruit. Second, the reflection effect (reflection effect), according to the reflection effect. In principle, the choice of profit prospects and the choice of loss prospects are equivalent, and the choice between the loss prospects is like a mirror reflection (a mirror image). Third, separation effect (isolation effect). In the usual case, the elimination method is not the only one. The diversity of the method of dealing with the problem can also lead to the disagreement between the preference and the choice. This is the so-called preference separation effect. And the psychological state often occurs when people make decisions: first, the anchoring effect, people are in the judgment. Second, regrets theory, psychologists and other psychologists believe that when people make mistakes, even a small mistake, there is a pain of regret, and a severe self remorse, not a further backdrop. Third, overconfidence (o Ver confidence) behavioral economics also found a phenomenon in the process of research, that is, people are overconfident of their own judgment. Finally, we give the expected theoretical model and point out that it is different from the expected utility theory model.Kahneman and Tversky. They are in its "expected theory: the risk decision". In the analysis (1979), another analytical framework is put forward.
WP=41
Theory is axiomatic, and their expectation theory is descriptive, derived from empirical observation in a deductive way, rather than in a group of logically attractive axioms. The detailed difference between the model is: first, in the expected theory, the decision maker does not care about the ultimate value of the wealth itself. The reference value of the heart wealth relative to a reference value, delta W., is often a wealth level of the decision-maker's current period, which means that the profit and loss are defined according to the status quo. The second point of expected utility is related to the value function v. A concave, which is convex for the loss, presents a decreasing trend of sensitivity to the change in the two direction. Moreover, it is bent on the origin. Compared to the case in which the right side of the origin is very small, the function is steeper in the left side of the origin when the loss is very small. On the contrary, the function u is usually smooth and concave function in the expectation efficiency theory. Third, the decision weight function Pi is the function of the objective probability P and Q. This function is monotonically increasing and discontinuous at 0 and 1. Therefore, for small probability, it always gives too large weight; and for large probability, a small weight is always given.
The fifth chapter makes an objective evaluation of the expected theory, and points out its shortcomings, and its substitution and role relationship with the expected utility theory. Finally, it points out its practical significance to our country. Although the construction of the basic theoretical framework of the expected theory remains to be perfected, it is more approaching the human behavior in the real world.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2004
【分类号】：F069.9

【引证文献】