I am analyzing a longitudinal panel dataset using OLS. The data spans around 40 years, but for some variables data was unavailable for certain categories. In most cases, the data for given category stretches not as far back as for others (or ends earlier), but there are also a few cases where a category has an entire blank series. Unfortunately there is no way that I can retrieve the missing data. Hence, I will have to estimate my models with gaps in my dataset.
I am wondering whether there are consequences for my estimation results that I should be aware of. I'm thinking here statistical biases or issues regarding the validity of my results. Suggestions would be very much appreciated.
