|
A condition in which the observations in a series are not independent of each other. The significance of auto-correlation for securities markets is that if today's price change is independent of tomorrow's price change in direction and magnitude, a variety of statistical techniques can be used to value securities and measure risk-but technical analysis would be demonstrably worthless. If the direction and magnitude of today's price change has implications for tomorrow's, the statistician must exercise more care in making inferences from the data-and the technician cannot be discredited.
|
