Abstract for

"What Does Composite Index Of NYSE Represent In The Long Run?"

Here we analyze the Composite Index (CI) for 35 years [From 1966 to 2000] of the New York Stock Exchange (NYSE) collected on daily-basis. The NYSE Composite Index averages the prices of all the stocks traded on the New York Stock Exchange. This index is automatically updated after each transaction and is sent electronically to the trading floor. We want to investigate what type of dynamics this data represents. There are plenty of works on the chaotic dynamics of the stock exchange data since the very beginning by Manelbortâ€™s pioneering work investigating the price changes dynamics of an open market in 1963. While the fluctuation of individual share prices following prevailing market relations makes them unpredictable, is it true for CI also? Particular question we ask is what dynamics of the stock exchange does CI represent? We find that up to first 16 years starting from the year 1966, CI is confined to values in the range of 36 to 75 and fluctuations are very small (20 points on either side). For the next 16 years that is up to year 1997, CI rises to 600 point mark although with fluctuations. For more precise analysis using nonlinear techniques, we break up the entire dataset in two parts- each having 16 years of CI points. To investigate the nonlinearity of the data, we resorted to the technique of surrogate data analysis [Das et al., Complexity Int., 2002:9].. We have calculated the Lyapunov Exponent (LE) of both the original and surrogate of each dataset and then compared the values. For the first half of CI data, the change is only 1.3% while for the next half, the change is 8.7%. The difference between a highly chaotic data and its surrogate counterpart should be much higher, for example, above 30% in our earlier work with electroencephalogram (EEG) data collected from human brains [Das et al., Complexity, 2002:7,3]. So it can be safely said that CI is not unpredictable in longer time scales. Moreover, the second half of the data fits well with some growing function of time- although with small random fluctuation predominant particularly in the last 3 years data. This confirms the observation that CI is a fairly good indicator of general market strength â€“ here, the US economy itself.