Why is intrapopulation genetic variality greater?

That's not exactly the case. What you do in a case like this is estabish the intrapopulation (you might say inherent) variability. Then you establish the amount of variability between the populations. You substract the intrapopulation variability from the interpopulation variability, and this is the amount of variability due to being from different populations (as opposed to the amount of variability due to just being a person).

More simply,
Let X=the average variability observed within the French and German populations.
Let Y=the observed variability of the pooled sample (French and Germans together).

Let Z=Y-X.

Z, then, is the amount of variability that is introduced by pooling the samples -- or, rather, it is the amount of variability between French and German populations. It's basic statistics, but it's not something most people have come across, so it's not readily apparent.

Incidentally, you don't have to be as close as French and German for this to be the case. In every ethnic pairing that's been examined, the intrapopulation variability is considerably greater than the variability between populations.

80% of genetic variability occurs within populations.