Network‐based early warning system to predict financial crisis

Abstract In recent years, early warning systems have been developed as a significant tool to prevent financial and economic crises. In this paper, we applied the forward‐looking conditional value at risk (COVaR) as a market‐based systemic risk measure to analyse the systemic risk in an emerging market. A network representation of asset exposures is presented based on the values of ΔCoVaR. According to the exposure networks, two new network‐based indices are introduced. Furthermore, regarding the dependence of the new indices on the values of forward‐looking CoVaR, they are applied as the indicators for the proposed early warning system to predict downturns and crises in the level of firms and the whole market. The results of the early warning system for the market return show that the indicators based on both proposed models have a good ability to predict the crises in the market. In other words, the indicators issued a warning signal up to seven periods before the real crisis in the market. Similarly, the results of the early warning system to predict downturns at the level of firms also indicate the ability of indicators to issue the true warning signals. The results of the proposed system can help policymakers and authorities to determine the appropriate microprudential and macroprudential policies to avoid systemic risks.