Behavioral and Network Origins of Wealth Inequality: Insights from a Virtual World

Almost universally, wealth is not distributed uniformly within societies or economies. Even though wealth data have been collected in various forms for centuries, the origins for the observed wealth-disparity and social inequality are not yet fully understood. Especially the impact and connections of human behavior on wealth could so far not be inferred from data. Here we study wealth data from the virtual economy of the massive multiplayer online game (MMOG) Pardus. This data not only contains every player's wealth at every point in time, but also all actions over a timespan of almost a decade. We find that wealth distributions in the virtual world are very similar to those in Western countries. In particular we find an approximate exponential distribution for low wealth levels and a power-law tail for high levels. The Gini index is found to be , which is close to the indices of many Western countries. We find that wealth-increase rates depend on the time when players entered the game. Players that entered the game early on tend to have remarkably higher wealth-increase rates than those who joined later. Studying the players' positions within their social networks, we find that the local position in the trade network is most relevant for wealth. Wealthy people have high in- and out-degrees in the trade network, relatively low nearest-neighbor degrees, and low clustering coefficients. Wealthy players have many mutual friendships and are socially well respected by others, but spend more time on business than on socializing. Wealthy players have few personal enemies, but show animosity towards players that behave as public enemies. We find that players that are not organized within social groups are significantly poorer on average. We observe that “political” status and wealth go hand in hand.

纵观全球各类社会与经济体，财富分配均不遵循均匀分布模式。尽管数个世纪以来学界已以多种形式收集财富相关数据，但现有观测到的财富差距与社会不平等现象的根源仍未得到完全阐释。尤为关键的是，迄今为止人类行为对财富的影响及其关联机制仍无法通过现有数据推导得出。本研究聚焦大型多人在线游戏（massive multiplayer online game, MMOG）《Pardus》中的虚拟经济体财富数据。该数据集不仅涵盖了每位玩家在任意时间点的财富状况，还完整记录了近十年间玩家的所有行为操作。研究发现，该虚拟世界中的财富分布与西方国家的财富分布模式高度相似：低财富水平下近似服从指数分布，高财富水平则呈现幂律尾部特征。研究测得其基尼系数（Gini index），该数值与诸多西方国家的基尼系数水平相近。我们还发现，玩家的财富增速与其加入游戏的时间存在关联：早期入驻的玩家，其财富增速显著高于后期加入的玩家。通过分析玩家在社交网络中的位置特征，我们发现贸易网络中的本地位置与玩家财富的关联性最强。高财富玩家在贸易网络中拥有较高的入度与出度，其邻居节点度相对较低，且聚类系数（clustering coefficient）偏低。高财富玩家拥有大量双向好友，在社交层面获得他人的高度认可，但他们在游戏中投入商业活动的时间远多于社交互动。高财富玩家鲜有私人仇敌，但会对被视为公敌的玩家表现出敌意。研究显示，未加入社交群组的玩家，其平均财富水平显著低于群组内玩家。我们还观察到，游戏中的"政治"地位与财富水平呈显著正相关。