Is having a maximum lag length of 0 ok after the differencing of variables (variables are not stationary at level). If yes, how can we run the Granger Causality Test through VAR or VECM modelling without having any lags, because it asks for minimum 1 lag to run the Granger Causality Test. The data is annual and I am trying to find the Causality relationship between GDP and Construction Industry.

Should I use VAR modelling or not when the number of lags is 0?
If not, which method should I use?

  • $\begingroup$ umer, what do you think about my answer? Is it clear or do you need further elaboration? (I see you have not accepted it.) $\endgroup$ Sep 7, 2017 at 13:10
  • $\begingroup$ umer, just as Hardy asked... do you need any clarifications associated with my answer. $\endgroup$
    – Sympa
    Nov 29, 2019 at 19:40

2 Answers 2


The case of no cointegration

This is easy. If you have no lags, then the model looks like \begin{aligned} \Delta x_{1,t} &= \gamma_{0,1} + u_{1,t}, \\ &\dots \\ \Delta x_{k,t} &= \gamma_{0,k} + u_{k,t}, \\ \end{aligned} where $k$ is the number of series in your model, $\gamma_0$s are intercepts (they would be set to zero if there are no time trends in the nondifferenced $x$s) and $u_t$s are error terms. Then clearly the history of series $j$ is not useful in predicting the series $i$ beyond the history of the series $i$ itself. (Actually, the history of series $j$ is not useful in predicting the series $i$, period.) And this holds for any $(i,j)=1,\dots,k$ where $i\neq j$. Therefore, none of the series Granger-causes any other series. (Also, no group of series Granger-causes another group of series.)

The case with cointegration

Consider a bivariate model for simplicity. Suppose \begin{aligned} \Delta x_{1,t} &= \gamma_{0,1} + \alpha_1 (x_{1,t-1}+\beta x_{2,t-1}) + u_{1,t}, \\ \Delta x_{2,t} &= \gamma_{0,2} + \alpha_2 (x_{1,t-1}+\beta x_{2,t-1}) + u_{2,t} \\ \end{aligned} $\beta\neq 0$ and either $\alpha_1\neq 0$ or $\alpha_2\neq 0$ or both. Then \begin{aligned} x_{1,t} &= \gamma_{0,1} + (\alpha_1+1) x_{1,t-1} + \alpha_1\ \beta x_{2,t-1} + u_{1,t}, \\ x_{2,t} &= \gamma_{0,2} + \alpha_2 x_{1,t-1} + (\alpha_2 \beta + 1) x_{2,t-1} + u_{2,t}. \\ \end{aligned} If $\alpha_1\beta\neq 0$ (i.e. if $\alpha_1\neq 0$ because we already know that $\beta\neq 0$) in the equation for $x_{1,t}$, $x_2$ Granger-causes $x_1$.
Also, if $\alpha_2\neq 0$ in the equation for $x_{2,t}$, $x_1$ Granger-causes $x_2$.

We also know that under cointegration there will be Granger causality at least one way (since $\beta\neq 0$ and either $\alpha_1\neq 0$ or $\alpha_2\neq 0$ or both), so either $x_1$ Granger-causes $x_2$ or $x_2$ Granger-causes $x_1$ or both.


The straightforward answer is if you have no lags, you have no VAR and no Granger Causality. By definition both methods require at least one lag. Respective definitions at Wikipedia are pretty helpful and clear on the concept.

The Present can be associated with the Present. That relationship may not have a determined direction (correlation, linear regression).

The Past cause or Granger cause the Present. That is where VAR and Granger Causality come in.

And, the Past is captured with the Lags. Without Lags you have no Past, no VAR, no Granger Causality.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.