摘要
This paper aims to both develop and apply advances from the field of complex networks to large economic systems and explore the (dis)similarities between economic systems and other real-world complex networks. For the first time, the nature and evolution of the Dutch economy are captured by means of a data set analysis that describes the monetary transactions among 105 economical activity clusters over the period 1987-2007. We propose to represent this data set as a multi-weighted network, called the monetary transaction network. Each node represents a unique activity cluster. Nodes are interconnected via monetary transactions. The millions of euros that traverse the links and that circulate inside each activity cluster are denoted by a link weight and a node weight respectively. By applying innovative methodologies from network theory, we observe important features of the monetary transaction network as well as its evolution: (a) Activity clusters with a large internal flow tend to cooperate with many other clusters via high volume monetary transactions. (b) Activity clusters with a lower internal transaction volume prefer to transact with fewer neighboring nodes that have a higher internal flow. (c) The node weights seem to follow a power law distribution. Surprisingly, (b) and (c) have been observed in community structures of many real-world networks as well. (d) Activity clusters tend to balance the monetary volume of their transactions with their neighbors, reflected by a positive link weight correlation around each node. This correlation becomes stronger over time while the number of links increases over time as well.
- 出版日期2011-10