A multi-layered network approach for the corporate and banking sectors: unwrapping the fundamental features of velocity of money
利用日本24年企业、供应商和银行关系的大数据集，构建多层网络计算货币流通速度，发现其下降主要由大公司持有更多存款和大型银行主导存款渠道导致，而非自由市场固有特征。
The declining trend in Velocity of Money, first observed within the Japanese economy but now affecting most of the advanced economies, is an important matter of concern to both central governments as well as financial market authorities. Yet, most of research to date has been top-down, and based on theoretical analyses and descriptions at the highest, countrywide, aggregation level. Inspired by the underlying principles of network science and graph theory, we propose a novel, extricated and disentangled method to compute the Velocity of Money at both economy-wide as well as at distinct coarse-grained, segmental levels. This is done by making use of a limited number of attributes from a large dataset of Japanese companies, suppliers and banking relationships, to build a multi-layered interconnected company money flow and bank payments networks over a period of 24 years. By applying our method, we are able to gain a much enhanced understanding of the micro features and dynamics of interactions among companies and banks as economic agents that shape the pace of money flows and accumulation of money stocks, the two major elements of the Velocity of Money. We find that the observed declining trend is not the result of a generic and inherently evolutionary feature of free market economies. Instead, in the case of Japan, the decline can be explained by two major aspects. First, the largest companies have proportionally reduced their average Velocity of Money (by holding higher levels of deposits) and second, larger banks became more dominant (by being overall better at canalising deposits). In addition, we find that the Velocity of Money is highly influenced by specific companies' attributes such as size, industry sector and geographical location with the highest values within the largest cities and prefectures.