Sometimes leads are included on the basis of the following reasoning: The future cannot cause its past, so if there is a relation between, say x[t+1] and y[t], it is not causal. Consequently it must reflect some type of non-causal relationship such as reverse causality or spuriousness. If that is the case it may raise questions about the nature of a relationship between x[t-1] and y[t].
- David Greenberg, Sociology Department, New York University
----- Original Message -----
From: Misha Spisok <[email protected]>
Date: Wednesday, September 30, 2009 8:47 pm
Subject: st: Why include leads (in a model with lags)?
To: [email protected]
> Hello, Statalist!
>
> My question in brief is, "What does including leads in a model that
> 'should' only have lags tell me?"
>
> My intuition is that if leads are statistically signficant, then
> something funky is going on (e.g., "How could future crime rates
> 'predict' current law enforcement expenditures?", or "How could future
> unemployment rates 'predict' current GDP growth?"). However,
> "something funky" is not intellectually satisfying.
>
> I am considering a model of the following form:
>
> y_t = b_0 + b_1*X_t + b_2*X_t-1 + b_3*X_t-2 + u_t
>
> It was suggested to me to use lead values of X, e.g., estimating a
> model of the form,
>
> y_t = b_0 + b_1*X_t + b_2*X_t-1 + b_3*X_t-2 + b_4*X_t+1 + b_5*X_t+2 +
> e_t
>
> If the coefficients b_4 or b_5 are statistically different from zero,
> should this be cause for alarm? If so, why?
>
> Is there a command in Stata (or other references, a tutorial, etc.)
> that can help me understand this (possible) problem better?
>
> Best,
>
> Misha
> Using Stata 10.1
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
