Description

Model history

Augmented Dickey-Fuller test (ADF-test) is a statistical test used to check the presence of a unit root in the autoregressive model and, therefore, to determine the stationarity of a time series. In financial markets, it is widely used to analyze time series of asset prices.

The test was developed and described by British statisticians David Dickie and Wayne Fuller in 1979. An extended version of the test (ADF) was proposed to improve accuracy and reliability over the original Dickey-Fuller test.

What the model is used for in financial markets

Stationarity test: The ADF test is used to determine whether a time series is stationary or trend-bearing (i.e., contains a unit root). Stationarity is important in time series modeling and econometric analysis because many methods require the data to be stationary.
Integration Analysis: Most often used in cointegration analysis, where multivariate time series are examined for long-run relationships.
Algorithmic Trading Solutions: Stationarity checking helps in the development of time series based strategies, including pair trading and time series predictions.

How the extended version is better than the regular version

Accounting for autocorrelation: The ADF test incorporates additional lags of differences in the dependent variable into the model to account for high autocorrelation in the time series data, making the test more robust and accurate.
Flexibility: The extended version allows both random walks and deterministic trends, including linear trend, to be considered in model estimation.
Reduction of autocorrelated residuals: the ADF test corrects the problem of autocorrelation of residuals, which can skew the results in the original version of the Dickey-Fuller test.

Dickey–Fuller test

Functionality

Widgets

The ADF model has 2 widget options: ADF and ADF Time Dominance.

ADF calculates the Dickey-Fuller test score and ranks assets with minimum/maximum ADF score.