基于PageRank的系統(tǒng)重要性金融機構識別模型
發(fā)布時間：2018-09-08 14:52
【摘要】：相對于依賴市場價格數(shù)據(jù)的標準計量統(tǒng)計方法,基于機構間雙邊敞口網(wǎng)絡拓撲結構的金融網(wǎng)絡模型更有助于系統(tǒng)重要性金融機構的識別和系統(tǒng)性風險評估。本文構建了貼近現(xiàn)實的CDS市場網(wǎng)絡模型,并基于單個違約機構傳染機制的分析,借鑒特征向量中心度和PageRank算法思想,研究建立了系統(tǒng)重要性金融機構識別的度量模型。本文所采用的排名技術算法在應對大規(guī)模金融網(wǎng)絡數(shù)據(jù)時具靈活性和可行性。測試結果顯示,監(jiān)管當局不僅要關注"太大而不能倒"的機構,更須將金融網(wǎng)絡中"關聯(lián)太緊密而不能倒"的中心節(jié)點作為問題認真加以對待。
[Abstract]:Compared with the standard statistical method which relies on the market price data, the financial network model based on the topological structure of the inter-agency bilateral exposure network is more helpful to identify systemically important financial institutions and to assess the systemic risk. In this paper, a realistic CDS market network model is constructed, and based on the analysis of the contagion mechanism of a single default institution, the measurement model of systemically important financial institution identification is established based on the feature vector centrality and PageRank algorithm. The ranking algorithm used in this paper is flexible and feasible in dealing with large scale financial network data. The test results show that regulators should not only focus on "too big to fail" institutions, but also take the central nodes of the financial network "too closely connected to fail" as a problem.
【作者單位】： 中南大學商學院;中國人民銀行鄭州培訓學院;
【基金】：國家自然科學基金資助項目(71173241;71473275) 教育部新世紀人才基金資助項目(NCET-10-0830)
【分類號】：F831.2
，
本文編號：2230846
[Abstract]:Compared with the standard statistical method which relies on the market price data, the financial network model based on the topological structure of the inter-agency bilateral exposure network is more helpful to identify systemically important financial institutions and to assess the systemic risk. In this paper, a realistic CDS market network model is constructed, and based on the analysis of the contagion mechanism of a single default institution, the measurement model of systemically important financial institution identification is established based on the feature vector centrality and PageRank algorithm. The ranking algorithm used in this paper is flexible and feasible in dealing with large scale financial network data. The test results show that regulators should not only focus on "too big to fail" institutions, but also take the central nodes of the financial network "too closely connected to fail" as a problem.
【作者單位】： 中南大學商學院;中國人民銀行鄭州培訓學院;
【基金】：國家自然科學基金資助項目(71173241;71473275) 教育部新世紀人才基金資助項目(NCET-10-0830)
【分類號】：F831.2
，
本文編號：2230846
本文鏈接：http://www.sikaile.net/jingjilunwen/huobiyinxinglunwen/2230846.html
教材專著
