Ivar function principle and difference with AR

Although there is a description about ivar, I don not understand the principle of it and the difference between it and AR. It seems the same, but not entirely.

Answer 1

[TARUN]

Hi @Sammy Zhang,

The AutoRegressive (AR) model and the ivar function are closely related but they differ significantly in the estimation methods used.

AR (AutoRegressive) Model forecasts future values of a time series based on its own past values. The model is expressed as:

A(q)y(t) = e(t) 

where A(q) is a polynomial in the lag operator q, y(t) is the time series, and e(t) is white noise.

ivar Function:

The ivar function estimates an AR model using the Instrumental Variable (IV) method.

Coming to your second question, some of the key differences between AR Model and ivar model are:

1. Estimation Method:

• AR Model is typically estimated using Maximum Likelihood when no endogeneity is assumed whereas ivar function utilizes the Instrumental Variable (IV) method to address potential endogeneity problems by using instruments.

2. Application:

• AR Model is most suitable when predictors are not correlated with the error term whereas ivar function is preferred when there is endogeneity.

You can learn more about these models here:

ivar: https://www.mathworks.com/help/ident/ref/ivar.html

AR: https://www.mathworks.com/help/ident/ref/ar.html