Thought this might be of interest: Brandon Berg, one of my co-conspirators at Catallarchy, wrote a post a while ago sketching out a possible way in which the decline in fertility rates might be overstated in the statistics that are commonly bandied about. The basic idea is that if the mean age of motherhood is increasing each year, the stats are likely to be skewed because of the way they're calculated (simple summing of the fertility rates for each age group). Now he's written a short follow-up demonstrating how sensitive the fertility numbers are to this effect. He suggests that the European fertility rates could realistically be understated by about 0.1-0.3, which is not huge but still significant.

Update: Related: Population Fallacies Part 2