Does composite index of NYSE represents chaos in the long time scale? Atin Das a, * , Pritha Das b a Naktala High School 1/257, Naktala, Kolkata 700 047, India b Department of Mathematics, BES University, Shibpur, Howrah 711 103, India Abstract Here we analyze the Composite Index (CI) for 32 years, from 1966 to 1997 of the New York Stock Exchange (NYSE). Individual share prices may be unpredictable—is it true for CI also—particularly in the time scale of 32 years? In the first half (consisting first 16 years) CI is confined to values in the range of 36–75 and in the second half, it rises to 600 point mark. Non-linear analysis of data confirms that CI is not unpredict- able in longer time scales. Moreover, the second half of the data fits well with some growing function of time. Ó 2005 Elsevier Inc. All rights reserved. Keywords: Financial time series; Non-linearity; Chaos; Lyapunov Exponent; Hurst exponent 1. Introduction Many systems in nature are self-similar when viewed at different scales. This property, known as ‘‘scaling’’, is characterized by a power law distribution (see 0096-3003/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.amc.2005.04.096 * Corresponding author. E-mail addresses: dasatin@yahoo.co.in (A. Das), prithadas01@yahoo.com (P. Das). Applied Mathematics and Computation 174 (2006) 483–489 www.elsevier.com/locate/amc