Inflation transmission diagnostics via a Bayesian graph vector autoregressive model with Markov switching

The dynamic transmission of inflation is a common phenomenon, and alleviating inflationary pressures is an important macroeconomic challenge for all countries. As a typical macroeconomic variable, the contemporaneous causality among inflation is not effectively examined, which may miss some useful policy information. Bayesian graph vector autoregression (BGVAR) model can identify contemporaneous causality and lagged causality among economic variables, but it lacks applied research on inflationary inflation. Considering the structural transformation in the inflation transmission process, this paper combines Markov switching (MS) with BGVAR and proposes a Bayesian graph vector autoregressive model with Markov switching (MS-BGVAR). The proposed model considers both regime switching and contemporaneous causality among macroeconomic variables. In this paper, simulation experiments are conducted to generate moderately dimensional simulated data under two regimes, and other indicators such as the identification accuracy of the graph structure show the theoretical reliability of our model. In addition, to demonstrate the validity of the proposed model on the economic domain, we choose inflation data from the Group of Seven (G7), which is representative of developed countries with inflation. We apply the proposed model to the identification of structural breaks in the inflation transmission process and causal transmission relationships in the G7 economies, and the experiments show that the proposed model has significant economic significance and good explanatory power in the selected target countries.