Cross-Correlation

Introduction: Cross-Correlation in Time Series Analysis

Cross-correlation is a statistical method used to measure the relationship between two different time series. It identifies whether changes in one series are related to changes in another, often at different time lags.

Background

While auto-correlation measures the internal dependency of one series with its past values, cross-correlation compares two distinct series. This makes it especially useful for detecting lead-lag relationships, where one variable may predict or influence another.

Key Elements / Features

• Two Series: Examines relationships between two time-dependent datasets.
• Lag Analysis: Measures correlations at multiple time lags to see if one series leads or follows the other.
• Direction and Strength: The cross-correlation function (CCF) shows whether the relationship is positive or negative, and how strong it is.

Applications / Examples

• Economics: Analysing whether changes in consumer spending follow changes in income.
• Finance: Studying if stock A’s price movements predict those of stock B.
• Operations: Checking if production volume influences later sales figures.

Example: A retailer may find that advertising spend (series A) is positively cross-correlated with sales (series B) at a lag of one week, meaning ads drive sales after a short delay.

Relevance / Impact

Cross-correlation helps reveal cause-and-effect patterns, detect delayed effects, and build predictive models. However, spurious correlations can occur if both series are driven by a common trend, so results should be interpreted with caution.

See also

• Auto-Correlation
• Time Series Analysis
• Forecasting
• ARIMA