Predicting Stock Index Changes

The addition or deletion of companies to/from a stock index has major consequences, with a change in demand from index-tracking investors and funds constituting the most obvious one. Numerous event studies since the 1980s have evidenced the existence of abnormal returns and trading volumes ('index effect') around the announcement and actual change dates for stock indices. This paper presents a model-driven approach to predicting changes in the index membership itself. Index rules are combined with models for the evolution of parameters such as market capitalization that drive a company's potential inclusion or exclusion. Special attention is paid to the inherent model risk. The 2021 revision of Germany's blue-chip and mid-cap indices, DAX and MDAX, both in terms of the number of index members and admission criteria, serves as a case study