<p>sakky,</p>
<p>That’s only so you can reject the null hypothesis, which is just the first step. After that, you can look at the coefficients and figure out the variance.</p>
<p>I’d think that you could do a somewhat interesting model with some dummies…</p>
<p>But either way, it won’t be a good model simply because there’s no way to do a good panel.</p>