Abstract: Following the principles and concepts of prudent macro management, a macro stress testing method for banking credit risk is put forward based on industry correlation. By considering industry correlation and the distribution characters of risk factor t, the multiple risk factor model is extended. The macro stress testing scenario with multiple risk factor model is considered, the values of industrial cycle index obtained under the stress scenario are transformed into conditional distributions of industrial risk factors under the corresponding scenario. Based on examination of macroeconomic cycle, the stress scenarios settings employ a variety of statistical methods to deal with historical macroeconomic data of the entire cycle in order to eliminate the procyclicality of credit risk measurement. The statistical methods include exponential smoothing, regression modeling approach and historical scenario analysis method. The process organically links the economic capital management with the prevention of systemic risk in banking industry. This macro stress testing method can reflect default correlations between credit assets in different industries, identify the negative impact on credit assets of other industries caused by some certain industry downturn, which reflects the source and mechanism of systemic risk.

Key words: macro stress test, industry correlation, systemic risk, procyclicality, monte Carlo simulation