Indexing Financial Stress

As we have experienced one of the greatest downturns in worldwide economy, there is a high demand for preventing future recessions. It is possible to translate disruptions in the nancial markets into nancial stress. We use several previously untried methods for constructing a nancial stress index. With these indexes, we are able to capture historical events in terms of nancial stress, such as the default of Lehman Brothers and the huge decrease in housing prices. By this, the evaluated nancial stress index seems useful, and more importantly, we can forecast high levels of stress in order to prevent future downturns. We nd that the clustering algorithm with dissimilarities based on Euclidean distances and clustered with the partitioning around medoids (PAM) algorithm performs the best. Overall, this paper shows how nancial stress can be indexed in various ways, and recommends that future research should account for nancial stress indexes based on clustering algorithms.