Cluster Analysis of Jumps on Capital Markets

Abstract

Cluster Analysis of Jumps on Capital MarketsWe analyze the behavior and performance of multiple price jump indicators across capital markets and over time. By using high-frequency we perform cluster analysis of price jump indicators that share similar properties in terms of their performance in that they minimize Type I and Type II errors. We show that clusters of price jump indicators do not exhibit equal size. Clusters are stable across stock market indices and time. Detected numbers of price jumps are also stable over time. The recent fi nancial crisis does not seem to aff ect the overall jumpiness of mature or emerging stock markets. Our results support the stress testing approach of the Basel III. Accords in that the jump component of the volatility process does not need to be treated separately for the purpose of stress testing.