How to run a regression with missing data?

I'm conducting a longitudinal study with approximately 28 observations/country. Unemployment data is missing for some (4-5) years. What I've done is that I have taken an average of the preceding and the following year. For Example, 2006=20% ; 2006=? ; 2008=25%. I've umed 2006's Unemployment to be 22.5%. Is it right? What do you suggest?