Under the neutral theory, the basic divergence formula is d = 2uT (divergence is equal to twice the mutation rate times time)

The factor of 2 comes from the fact that when comparing 2 sequences that diverged time T in the past, there are 2 branches on which independent evolutionary changes can occur.

    X       |
   X X      |
  X   X     | T
 X     X    |
X       X   |

Rearranging this expression gives an estimate of the divergence time:

T = d/2u

This assumes a molecular clock, which effectively implies neutral sequence divergence and a constant mutation rate u. The reason that no explicit method is given is because this is (for a point estimates) a simple calculation.

Mutation rates can be estimated directly via mutation accumulation experiments or by measuring new mutations in parent-offspring comparisons. A good review on the topic can be found here: http://rstb.royalsocietypublishing.org/content/365/1544/1169.full

If you are interested dating the age of insertions using intra-element LTR-LTR comparisons, I suggest you look into using REannotate, which can perform the calculation of d (its standard deviation) automatically: http://www.bioinformatics.org/reannotate/