Региональная взаимозависимость индустрий

Interdependence of industries has always been of particular interest to researchers. The vast majority of the models that attempt to incorporate and/or quantify these interdependencies require a detailed information on trade flows and input-output tables that are constructed in a systematic and comparable across regions ways. Moreover, CGE models are based on abstract relations that are not observable directly and are subject to researchers subjectivity. Recognising these problems leads us to suggest copula approach for modelling dependencies of industries. This paper applies the D-vine copula based quantile regression analysis method in an attempt to identify relative importance of industries within a region and of countries within a given industry. We tailor D-vine copulas based on time series of averaged probabilities of default of companies in each industry and estimate the model non-parametrically. As a result, without complex and hard to obtain global input-output table, we can customise our proposed model to include variables of interest, i.e. specify industries and regions, in order to assess the risks of externalities and spillovers.