Трансмиссия системного риска между банковскими системами стран Азиатско-Тихоокеанского региона и России

Финансы: теория и практика. 2023. Т. 27. № 5. С. 182–194.

Dzuba S. A., Тишковец В. С., Shchepeleva M.

The subject of this research is systemic risk transmission between financial sectors in the international financial market. The purpose of our paper is to determine topology characteristics for the network connecting banking systems in the Asia-Pacific region (APR) and Russia. Given the growing role of this region in the global financial market, its susceptibility to crises can be dangerous for other countries. This determines the relevance of our study. To build the network, we used the SRISK indicators, which reflect capital losses in the financial institutions’ capital losses in case of a large-scale crisis. The networks were built with the use of the NETS algorithm, proposed by Barigozzi, M., & Brownlees, C. (2019). This method is based on sparse vector autoregressions estimated by LASSO. As a result of the application the algorithm, we get two networks —simultaneous interconnections and using the values of the lagged variables. The networks were constructed for the 2005–2020 time period and separately for sub-periods including the global financial crisis (2005–2013) and the COVID-19 pandemic period (2014–2020). Based on the results obtained, the networks over the entire time period seem to be quite susceptible to external risks. China, Japan, Singapore and Taiwan are the largest shock donors in this region. Russia mainly accepts risks, generated by other countries, in the period 2014–2020. Strengthened/weakened cooperation with the largest risk exporters in this region will increase/decrease the likelihood of systemic risk transfer to the Russian financial sector.

 

 

Keywords: сетевой анализ, network analysis, Granger Causality Test, centrality, центральность, network topology, systemic risk in the financial sector, sparse vector autoregressions, системный риск финансового сектора, разреженные векторные авторегрессии, тест Гранжера, топология сетевого взаимодействия