Interpreting Measured Serial Correlation In Univariate Time Series Analysis, With An Example From The New York Stock Exchange

Abstract

This paper draws attention to the divergence of the observed serial correlations, for a typical given time series, from the corresponding theoretical autocorrelation pattern of the underlying generating process, which is not just a matter of sampling variation, and which seems to be commonly overlooked when analysing data infinance. We illustrate our thesis with examples from a case-study in the recent literature, where stock market transactions data for two distinct periods were analysed, and the behaviour of returns and various trading characteristics examined.

Résumé

On veut signaler dans cet article la différence qui existe, dans le cadre d’une série chronologique typique, entre les corrélations sérielles observées et la structure correspondante des autocorrélations théoriques générées par le processus sous-jacent. Cette difference ne se résume pas à une simple variation échantillonnale; elle semble d’autre part avoir été Ignorée lors de l’analyse de données financières. En guise d’exemples, on se sert d’études récentes portant sur des transactions boursières effectuées lors de deux périodes distinctes; les données sont d’abord analysées, puis le comportement des taux de rendement et de certains indices boursiers est étudié.

Key Words:

• ARIMA Processes
• Box-Jenkins Modelling
• Long-Memory Processes
• Model Identification
• Nonstationary and Nearly Nonstationary Behaviour
• Serial Dependence
• Simplifying Operators

Additional information

Notes on contributors

Oliver D. Anderson

Oliver Anderson is a graduate of Cambridge, London, Birmingham, and Nottingham; with first class honours degrees in mathematics and economics, master’s degrees in mathematics and statistics, and a PhD in mathematical statistics. He has worked in industry, for government, and as a consultant statistician; and taught in schools, colleges and universities. He has lectured in over 20 countries and published some 350 items (including a book, 240 refereed papers and 70 other invited contributions), and has edited several journals and a further 15 books. He is an active member of a dozen professional societies (in England and abroad) concerned with Education, Training, Computing, Information Technology, Mathematics, Statistics, Operations Research, and Econometrics; and is a Fellow of the British Institute of Management. In 1979, he was honoured by election to the International Statistical Institute; and, in 1988, he was elected a Fellow of the American Statistical Association.