Abstract: There are many relationships among enterprises in the socio-economic system, such as credit relationships and guarantees relationships among enterprises, which forms complex connections among enterprises. Theses complex connections bring many economic benefits to the development of enterprises, butthey also provide a medium for risk contagion among enterprises.Network theory provides a new research perspective for the inter-enterprise correlations. Based on actual data, it has found that there are some typical network structural characteristics among enterprises, such as scale-free networks.Revealing the micro-mechanism of the formation of enterprise networks helps to understand their structural characteristics. Therefore, in this paper, the upstream and downstream enterprises are considered, and an endogenous network model is constructed to study the internal formation mechanism and evolvement characteristics of inter-enterprise credit connections. Through simulation analysis of the endogenous network model, the results show that the model constructed in this paper reproduces the characteristics of the enterprise system in real life:the degree distribution of enterprise credit network is a power-law distribution, and this means the network is a scale-free network; longer aggregation periods might be preferable to have a better description of inter-enterprise credit relationships, and the network density increases with the increase of the aggregation period; the distribution of enterprise assets has the power-law tail, and the asset growth rate of enterprises gradually converges with time, and its probability distribution is approximate to the normal distribution.The key research of this paper is to explain the formation mechanism of enterprise credit network, which is one of the main contents of network theory research. Moreover, this study is the basis for further research on related practical issues, such as constructing enterprise evolution systems based on network theory to study the effect of economic policies.

Key words: credit relationship, network model, scale-free network, network density