Unit Root

A unit root is a feature of some stochastic processes that causes shocks to have a permanent effect on the value of the series. If a time series has a unit root, it is non-stationary and its variance grows over time, which can lead to misleading statistical results.

In financial econometrics, testing for a unit root is a prerequisite for many types of modeling. A series with a unit root is often referred to as an integrated process.

Most raw financial price series contain a unit root, which is why they must be differenced before analysis. Failing to identify a unit root can result in spurious regressions where variables appear related when they are not.

Economists use tests like the Augmented Dickey-Fuller test to detect the presence of unit roots. Once a unit root is identified, the series can be transformed to achieve stationarity, allowing for accurate model estimation.

It is a fundamental concept for anyone working with time series data.