Year dummies in Fixed Effects estimation with panel data

Quick question. I'm working with an unbalanced international dataset. I'm using fixed effects because (at least mostly because) I want to allow for unobserved country specific effects.

My question is regarding using dummies for each year in my equation. Much of the literature on my subject doesn't use the year dummies, but I've seen some things saying it's good to use year dummies in FE estimation to allow for different intercepts for each year.

What would be the best reasons to either use or not use year dummies?

I'm hoping this will help me a) decide whether or not to use year dummies, and b) justify my decision, either way.