We operationalize the definition of systemic risk provided by the IMF, BIS, and FSB and derive a two-stage hierarchical hypothesis test to identify indicators of systemic risk. Applying the framework to a set of candidate variables for 45 countries, we detect two credit-based financial cycle variables that, by and large, pass our test. However, for many other variables, including the Basel III credit-to-GDP gap, we find that elevated systemic risk is signaled by high values in some countries and by low values in others. More generally, our results suggest that, ex ante, systemic risk can be clearly identified only once the turning points of indicators have been observed.

我们将 IMF、BIS 和 FSB 提供的系统性风险定义操作化，并推导出两阶段分层假设检验来识别系统性风险指标。将该框架应用于 45 个国家/地区的一组候选变量，我们检测到两个基于信用的金融周期变量，大体上通过了我们的测试。然而，对于许多其他变量，包括巴塞尔协议 III 的信贷与 GDP 的差距，我们发现系统性风险升高的信号是一些国家的高值和其他国家的低值。更一般地说，我们的结果表明，只有在观察到指标的转折点后，才能在事前明确识别系统性风险。