Record Statistics of Equities and Market Indices

Abstract

Record events in a time series denotes those events whose magnitude is the largest or smallest amongst all the events until any time N. Record statistics is emerging as another statistical tool to understand and characterise properties of time series. The study of records in uncorrelated time series dates back to 60 years while that for correlated time series is beginning to receive research attention now. Most of these investigations are aided by the applications in finance and climate related studies, primarily due to relatively easy availability of long measured time series data. Record statistics in respect of empirical financial time series data has begun to attract attention recently. In this work, we first review some of the results related to record statistics of random walks and its application to stock market data. Finally, we also show through the analysis of empirical data that for the market indices too the distribution of intervals between record events follow a power law with exponent lying the range 1.5–2.0.