Uri Gneezy Deception: The role of consequences[J]. American Economic Review, 2005, 95(1): 384-394.

实验综述

Approaches to deception

• moral approach (St. Augustine, 421).

  • In literature
  • 人会绝对考虑对方的利益
  • 欺骗会对个体造成不可接受的成本
  • “what a person should do.”

• economy approach

  • in economy
  • 自私、精于算计
  • 绝不考虑他人的利益
  • 欺骗不会对自我造成成本
  • “what a rational eco­ nomic agent should do.”

• mechanism design approach

  • 只有在激励相容(incentive-compatible)下才说实话
  • e.g.逃税：衡量说谎的收益（逃税额）和说谎的伴随成本（被发现概率和惩罚）

• consequentialist approach: weigh benefits against harm, and happiness against unhappines

  • by utilitarian philosophers
  • 说谎的行为被视为中性的：如果说谎和不说谎都能够达到相同的结果，那么两种途径是平等的

Classification of Lies, and a Definition

• category 1: help both sides

  • white lies: “You look great today!”

• category 2:help the other person even if it harms the liar

  • pure altruism (Becker, 1976)
  • impure motive according to which people enjoy the act of giving (James Andreoni, 1990)
  • efficiency motive according to which people prefer outcomes that enlarge total surplus (Gary Chamess and Matthew Rabin, 2002)

• category 3: do not help the liar but can harm both side

  • 可能是对不公平行为的恶意反应

• category 4: increase the payoff to the liar and decrease the payoff to the other party

  • in most economic events
  • relationship between the deceiver’s gain and the other’s loss.

• 研究问题：证明欺骗行为也受到外部性(Externality)的影响

  • 欺骗行为中的个人收益和他人损失带来的影响

• 研究方法

  • 反证法
  • 通过改变谎言的绝对和相对结果，测量撒谎倾向的变化

实验设计

A. Cheap Talk Sender-Receiver Game

• 基准实验:独裁者实验

  • 实验组成：sender A, receiver B ; message1,2 ；decision

  • Sender A

    • 获悉收益的分配方式，选择message给B

    

    • 部分A被要求猜测B的反应
      • 根据准确性付费

  • Receiver B

    • 选择接受/拒绝

  • Treatment：收益的分配方式不同

    

    • 对于A来说，B选项总是有利可图
    • 不同的treatment，代表不同程度的外部性

  • 实验结果

    • 不同的treatment下分布不同

      

      • 分配1 B巨亏A小赚：0.17
      • 分配2 B小亏A小赚：0.38
      • 分配3 B巨亏A巨赚：0.52
      • 如果决策者只关心自己的收益那么无论对方损失多少都不会改变A的决策，但事实不同的是，对A同样小赚的分配1,2，A的选择显著不同
• Control treatment：

  • 实验步骤（DG）

    • A直接决定分配，不需要B选择
    • 因为前面的实验中80%的A说谎，因此在本实验中A的决定只有80%的机会生效，这样能够有一致的预期回报
    • 只要DG中的欺骗率比Deception中更高，说明存在lying aversion抑制了欺骗行为

  • 实验结果

    

    

    • 所有的选择频率都大幅上升
    • 当A的选择性质从欺骗转为独裁后，作出同样选择的频率大大增加，这表明人们不仅因为在乎他人而激发诚实行为，而且会厌恶撒谎，人们的选择反映了非结果主义的偏好

B. 调查问卷

• 考察
  • 人们对于后果在谎言中的作用的认识
  • 对于不同谎言之间的相对公平的认识

“Mr. Johnson is about to close a deal and sell his car for $1,200. The engine’s oil­ pump does not work well, and Mr. John­ son knows that if the buyer learns about this, he will have to reduce the price by $250 (the cost of fixing the pump). If Mr. Johnson doesn’t tell the buyer, the engine will overheat on the first hot day, resulting in damages of $250 for the buyer. Being winter, the only way the buyer can learn about this now is if Mr. Johnson were to tell him. Otherwise, the buyer will learn about it only on the next hot day. Mr. Johnson chose not to tell the buyer about the problems with the oil pump. In your opinion, Mr. Johnson’s behavior is (please circle one): completely fair; fair; unfair; very unfair. What would your answer be if the cost of fixing the damage for the buyer in case Mr. Johnson does not tell him is $1,000 instead of $250? Mr. Johnson’s behavior is (please circle one): completely fair; fair; unfair; very unfair.”

• 作出以下场景的判断
• 实验成果

实验结果汇总

• 人们在不受惩罚情况下的说谎行为会考虑谎言的后果
• 人们的对于欺骗的认知是基于后果的

创新及意义

• 挑战了传统对于人的假设，提出影响欺骗行为模式的新因素
• 实验非常简单易懂
• idea：但是衡量外部性成本的方式有点少