S/F/38. Time series analysis of a banking sector index

This paper is dedicated to the analysis of the time-series behaviour of the UK bank industry stock index. The data sample for research covers the period from 12/1993 to 12/2004. The question under consideration is whether banking stock index can be regarded as market efficient. First of all, the test of random walk hypothesis is applied. Since volatility clustering is typical in financial data, the random walk hypothesis with increments is tested. The time-series tests reject the RW hypothesis. However, when the variance ratio tests are applied, the RW hypothesis cannot be rejected. This gives rise to the conclusion that even if the banking industry is not market efficient it is very close to this. The paper continues with testing ARMA and GARCH models to capture the behaviour of the index. The best models are chosen based on the results of appropriate tests. This work can be very useful for students doing univariate time-series analysis. It contains a very readable and understandable explanation of all the tests implemented and clear logic behind all conclusions. Thus, this paper is highly recommended for those who study time-series analysis and conduct practical tests of market efficiency.