联系我们

论文查询结果	返回搜索

论文编号：	6238
作者编号：	1120100796
上传时间：	2014/6/6 10:29:40
中文题目：	集体行动的实验研究 ——偏好、理性、目标与治理
英文题目：	An Experimental Study of Collective Action: Preference, Rationality , Target and Governance
指导老师：	李建标教授
中文关键字：	集体行动；偏好结构；理性；实验；治理
英文关键字：	Collective Action; Preference Structure; Rationality; Experiment; Governance
中文摘要：	集体行动是N个人在同一的博弈框架中，为实现共同的目标而采取的行动。本文聚焦于不同目标的集体行动的微观动因研究，构建了研究集体行动的一个理论框架，根据集体行动的目标，将集体行动分为集团福利主导型、社会福利主导型和他人福利主导型三类；关于集体行动的微观动因，我们关注个体偏好和理性两个维度，对偏好和理性范畴作了扩展。在偏好方面，我们认为个体的偏好是多维度偏好耦合的，个体既具有纯自利偏好，又具有“不纯的”社会偏好，还具有纯粹的社会偏好，个体的偏好结构调整是在一定的框架中进行的。在理性方面，我们提出远见假设，认为集体行动中的部分个体可以预见到博弈的各种均衡结果，可以意识到合作比不合作更能提高集体和自身的实际得益。如果他们认为群体中持有类似认知的个体足够多，合作的概率较高，则他们有可能选择合作而不是搭便车。 在集团福利主导型集体行动研究中，主要聚焦于经典的线性公共物品自愿供给博弈框架。在社会福利主导型集体行动研究中，关注其不同的层次，对一阶社会福利主导型集体行动的理论分析聚焦于集体抗争博弈和先驱者博弈。对二阶社会福利主导型集体行动主要关注个体建立私序来治理一阶集体行动的过程，以及社会规范在外部冲击下的变迁。 在理论分析的基础上本文进行了实验研究。在集团福利主导型集体行动方面，实施了强调远见的单期公共物品实验和多轮重复且重启的公共物品实验。在一阶社会福利主导型集体行动方面，实施了提议者提出分配方案、回应者可接受或反对的集体抗争实验，加入了领导权、序贯不对称、更小的人群规模和分配信息对称等设置。实施了先驱者博弈实验，设计了不同的集体行动外生难度参数、静态与动态博弈过程以及损失和得益框架。在二阶社会福利主导型集体行动方面，实施了社会规范构建与变迁实验，被试可以在两类惩罚机制间进行投票选择，随后进行公共物品供给实验。实验中被试对公共物品价值异质，而公共物品的价值可能外生地降低，已经建立的社会规范也可能外生移除。 主要的研究结论如下： 第一，理论分析表明，具有自利和社会偏好耦合的偏好结构的个体更有可能做出贡献，其贡献取决于公共物品的价值、社会偏好维度权重、不平等厌恶强度系数等。测度远见的实验结果表明，先强调远见的设置下个体单期合作水平显著提升，但后强调远见设置下个体的合作水平未见提升。对实验后调查结果的分析表明存在一定比例的具有远见、且真实地作出了较高贡献的被试。因此，远见可以导致个体理性地做出贡献。多轮重复实验结果支持了经验、远见和互惠假说，对多轮重复实验的回归分析表明，长期内个体表现出有条件的合作。总体上看，个体在单期的合作源于利他、不平等厌恶、暖意和远见理性，而长期内的合作更多是互惠驱动的有条件合作。 第二，对一阶社会福利主导型集体行动的理论分析表明，在参与者纯自利假设下，集体抗争博弈中的提议者可以成功实施DAC（Dividend and Conquer，分而治之）策略，使得回应者的抗争一直不会发生。但回应者的偏好、理性和博弈结构均可以导致集体行动。实验分析表明，集体抗争博弈中，提议者更多表现出自利偏好，他们主要采用DAC策略。公开提议者对所有人的分配信息有效地协调了抗争，此时提议者采用了更均等的分配。领导者会作出抗争的建议，序贯不对称的设置下更不易于发生集体抗争。在先驱者博弈中，外生难度显著影响个体参与集体行动的顺序，形成了发起集体行动的若干行动集团。 第三，二阶社会福利主导型集体行动的理论分析认为，对公共物品赋予更高价值（可能是出于社会偏好或远见理性）的个体会更愿意选择更有效的社会规范。当公共物品的价值降低，或更有效的社会规范因为某种原因而自动消失时，社会规范会发生变迁。但是，由于认知成本和既有规范的可自我实施性，社会规范的变迁表现出路径依赖。实验结果发现被试最终选择经过改进的、可自我实施且帕累托占优的社会规范，并且外部冲击下被试仍然从信念和实际选择两个层次上坚持既有社会规范，使得既有社会规范表现出路径依赖。在社会规范选择方面，被试主要考虑惩罚机制的效率。 本文创新点主要有四点：第一，构建了集体行动的一个统一解释框架并进行分类。第二，率先提出偏好结构假设和基于远见的行为理性假设并基于此解释集体行动的合作。第三，将集体抗争博弈扩展到多人序贯情况，并率先提出先驱者博弈。第四，率先设计和实施社会规范及其信念的变迁实验。当然，本文也有以下有待扩展之处：首先，在实验的样本选择和样本规模方面存在局限；其次，偏好结构缺乏公理性基础；第三，对信念变迁的研究还缺乏基于博弈学习理论和演化方法的刻画，这需要今后进一步的研究。 本文的实践意义在于为集体行动的社会治理和政策制定提供理论借鉴。随着经济的发展、社会利益主体与诉求多元化，现代社会必然进入集体行动蓬勃发展的阶段，这种趋势在我国已经显现。而成功的集体行动治理经验表明，对不同诉求的集体行动进行分类治理，对诉求和手段合理的集体行动制度化、规范化，对诉求和手段不合理的集体行动进行疏导和压制，是健全现代社会的主要途径。本文的研究为有针对性的集体行动治理提供理论支持，在建立集体行动预警指标体系和疏导方案、将集体行动导向制度化的道路上，本文的研究亦可提供借鉴。
英文摘要：	COLLECTIVE action is the action with N persons who share the same framwork of the game and want to achieve the same target. This dissertation focuses on the micro motivation and the category of the collective actions. We have built a theoritical framwork to explain the collecive action and classified it into three types based on their targets, that were Group Wellfare Leading, Social Wellfare Leading and Others’ Wellfare Leading (GWL, SWL and OWL for short). We assumpt that the mirco motivation can be summarized as two factors, which are preference and rationality, then we expand the definition of them.For preference, we assumpt the participants of collective action games have individual preference structures, which are the combination of three dimensions: selfish preference, impure social preference and pure social preference. The individual preference will adapt the situation by changing its structure basing on the framwork of decision making. For rationality, we assumpt that part of the participants have strategy foresight, they can realize the Pareto-dominated result of the use of dominate equlibrium in which people choose not to cooperate, then they will try to coorprate rather than free-riding to achieve a Pareto efficient result if they expect there will be enough farsighted ones. Based on these assumptions, we analyze several types of collective action. For the GWL, we mainly studied volunteer contribution mechniasm(VCM) for the linear public goods. Two kinds of first order SWL games are studied, that were collective resistence games and pioneer games. We focused on the private ordering governance of first order collective action and the change of social norms, which were the main content of the second order SWL games. We did the experimental studies based on the theoritical analysis. For the GWL games, the one-shot public good games had been did which highlighted foresight and repeated public good games with restart. For the first order SWL games, we run the collective resistence experiments in which the proposer could make a distribution and the recievers could choose to accept or reject. We designed the treatments to explore the effects of leadship, sequential and asymetric decision making, smaller group size and information symetric,etc. On the experimental study of pioneer games, we designed parameters of the exogenous difficulty of collective action, the static and dynamic procedur and gain-loss decision making framwork. For the second order SWL games, we implemented public good experiments with voting for alternative sanction mechanisms. In the experiments, subjects had heterogeneous value of public good, which could be lowed, and the established social norms could be removed exogenously. The main conclusions of this dissertattion are as follows. Firstly, theoretical analysis show that，the person with a preference structure which is combinated by selfish and social preference will be more likely to make contributions. The money she puts into the public account dependents on the value of public goods, the weight of the social preference and the strength of inequality aversion. The experimental results show that, the individual contribution of one-shot VCM games would be significantly higher in the first play than the second play in the treatment in which foresight was emphasized in the first play. However it is not ture in the game in which the foresight was emphasized in the second play. The questionare results show part of the subjects are farsighted ones, who make truely higher contribution.So the person with foresight will make contribution, which is in accordence with rationality. The results of repeated games support hypothesis of experience, foresight and reciprocity. The regression results confirmed the conditional cooperation. Generally speaking, in the one-shot games, the micro motivation of cooperation are altruism, inequality aversion and foresight. However, reciprocity drives the cooperation in long-run games. Secondary, theoretical analysis show that，if all the players are selfish, the proposer will use the DAC(Dividend and Conquer) strategy succesfully, at which time the resistence will not happen. However, the social preference, forsight and some features of the games can totally lead to collective action. In the experiments, the proposers are mainly selfish ones who used DAC strategy. Disclosure of the ditrisbution made by the proposers would coordinate collective resistence effectively, Expecting this, the proposer would turn to more equally distribution. The leaders prefered to advise the responders to resist in the treatment with leadship, while there were less collective resistence in the sequential asymetric treatments. In the pioneer games, the exogenous difficulty significantly influence the order which the players choose to participate in the movement. Some movement group had been built. Thirdly, the model of second order social wellfare orientation collective action argued that, the higher the participants’ value of the public goods are, the more efficient social norms they will choose. When the value goes down or the more efficient social norms disappeared exogenously, the social norms will change. A the same time, a path dependence will happen because of cognition cost and the self-enhance features of the existing norms. The experimental results found that the subjects finally choose the self-enhance and Pareto efficient so-cial norms. They insisted on choosing these norms and holding the belief that the efficient norms would be chosen even under the external shocks. Choosing the norms, the players mainly considered the efficiency of sanction. The main contributions of this dissertation may be as follows: (a) we created a general framwork to explain why and how the collective action happens, and classified the collective actions; (b)we developed the preference structure and foresight assumptions, explained the collective actions based on these; (c)we extended the collective resistence games to a multiple player and sequential situation. We also created a pioneer game; (d)we run the social norms and belief changing experiments. However, there are also some disadvantages of this dissertation: the sample choosing and size, the axiom base of the preference structure theory, the lack of games learning and the evolutional approach, etc. We will consider these in our later researches. The practical meaning of this research is the reference value to the social governance and policy making about collective actions. While the economy develops quickly and people’s request turns to be multiplex, the collective action is preseting to be normal in morden society, including China. The sucessful governance experience of collective action shows that, the collective action should be treated differently. We should institutionalize and normalize the collective action with reasonable request and method, and dredge and suppress the the collective action with unreasonable request and method. This dissertation supports the governance of collective action theoritically, and supplies some references about the early-warning system and resolutions of the collective actions.
查看全文：	预览  下载(下载需要进行登录)