Financial Crisis Early Warning Based on Panel Data and Dynamic Dual Choice Model

Based on the research of currency crisis pressure index, bank crisis pressure index, and asset bubble crisis pressure index, this paper introduces an external shock pressure index reflecting the impact of global economic changes on economy and synthesizes systemic financial crisis pressure based on the above four pressure indexes; then, all the alternative early warning indicators and the systemic risk pressure index constructed in this paper were tested for Granger causality. We build financial systemic risk pressure indexes, including currency crisis pressure (CCP) banking crisis pressure (BCP) index, bubble crisis pressure (PBP) index, and external shock pressure (ESP) index to predict financial crises. Finally, four indicators that have a significant impact on the systemic financial crisis pressure index were selected, namely, the stock price index change rate, industrial added value growth rate, domestic and foreign real deposit interest rate differential, and foreign direct investment as a percentage of GDP. A dynamic Logit model with lagging binary variables is constructed, and compared with the traditional static Logit line, the actual dynamic fitting effect is better than the static Logit model. The dynamic Logit model is used to predict the early warning status of systemic financial crisis in 2020, and the forecast of various early warning indicators is realized by the ARIMA model. The final prediction results show that the probability of a systemic financial crisis in China in 2020 is extremely low, almost zero. This is in line with the overall improvement in the international economic situation in 2020 and the steady growth of the domestic economy.