Cointegration is a statistical property of a collection of time series variables which indicates that a linear combination of them behaves like a stationary series, even though the individual series themselves are non-stationary. In simpler terms, two or more non-stationary time series can be said to be cointegrated if they share a common stochastic trend. This is crucial in econometrics, as it implies a long-term equilibrium relationship despite short-term fluctuations.
To determine if two series and are cointegrated, we can use the Engle-Granger two-step method. First, we regress on to obtain the residuals . Next, we test these residuals for stationarity using methods like the Augmented Dickey-Fuller test. If the residuals are stationary, we conclude that and are cointegrated, indicating a meaningful relationship that can be exploited for forecasting or economic modeling.
Start your personalized study experience with acemate today. Sign up for free and find summaries and mock exams for your university.