Autocorrelation, Serial Correlation, Lagged Correlation, Defined
The relationship between a given security and itself over various time intervals. It could be between a given time series and a lagged version of itself over successive time intervals. Serial correlations are often found in repeating patterns when the level of a security effects its future level. Serial correlation is used by technical analysts to determine how well the past price predicts the future price.
Autocorrelation measures the tendency for an above average return (momentum) to be followed by another above average return.
The above average return can be on the downside. For example, the waterfall decline in stock, bond, and commodity prices in the fall of 2008 was serial correlation. That is, prices continued to cascade down because past prices were falling. Panicked investors sold their positions because they were falling farther than they could tolerate. Contagion, such as spreading effect of fear, led to serial correlation.
From Alexander M. Ineichen. Absolute Returns: The Risk and Opportunities of Hedge Fund Investing (Wiley Finance):
"Serial correlation is also referred to as autocorrelation and indicates a condition in which the random error terms in a sequence of observations are not independent (violation of serial independence).
Alexander M. Ineichen. Absolute Returns: The Risk and Opportunities of Hedge Fund Investing (Wiley Finance) (Kindle Locations 5861-5862). Kindle Edition.
Serial correlation is also referred to as autocorrelation and indicates a condition in which the random error terms in a sequence of observations are not independent (violation of serial independence).
In other words, the daily price action is serially correlated if the daily action is correlated. Investors sometimes sell stocks for the simple reason they are falling sharply in recent days. Or, they may buy stocks because they are gapping up in recent days. Prices, therefore, are not random and not always disconnected or independent or random.
