I'm doing a fixed effects (cross-sectional fixed effects) estimation and since fixed effects estimations discard the information in cross-sectional differences by demeaning the data, I checked the correlations between the fixed effects and the average of my explanatory variables. I found that the estimated fixed effects are significantly correlated with one of the explanatory variables, let's call it x1. More precisely, I calculated, for each cross-section, the average value for x1 over time. It is this over-time-average of x1 that is correlated with the fixed effect. So I wondered: doesn't this correlation provide extra information? Intuitively, I would conclude that this correlation can be interpreted as: cross-sections with an on average higher value for x1, tend to have, on average, a higher value for y (the dependent variable). Nevertheless I cannot find confirmation for this intuition. Is this intuition correct? If not, are there any references or explanations that clarify why it is incorrect?
For econometric discussions not necessarily related to EViews.
1 post • Page 1 of 1
Who is online
Users browsing this forum: Google [Bot] and 15 guests