Cross-Section with Mismatched Years (NOT panel/pooled data)

[cuscus]

Hello all,

I am trying to estimate a cross-section model (across different countries), but some of the independent variables are only available for certain (but different) years.

The dependent variable and my independent variable that I'm primarily interested in have observations for each country every year. However, some of the other independent variables have observations in different years (but not like in an unbalanced panel). For example, variable A is observed for all countries in 2005, whereas variable B is only observed in 2007, and variable C is has some observed in 2005, and some in 2007, and variable D is only observed in 2006. Some of these variables measure social/infrastructure development, such as % with access to clean water, where another might measure % of population that smokes. That being said, I don't expect these variables to vary too much over a 3-year period. My question is whether or not it's acceptable to consider my cross-section time period to be 2005-2007, and just take the average of the data with annual observations.

I'm not looking for someone to drop an answer in my lap, as I'm sure this problem can be addressed in multiple ways. I'm just wondering if you guys are aware of any studies that have had the same problem that I can use for guidance. I've looked in my (undergrad) econometrics textbook and various other sources but haven't found anything about this problem.