While I am sure that estimates of weights of components designed but not yet made are getting better all the time we should not forget the effects of human nature.
Design team individuals responsible for one small part of the whole have been known to over estimate the weight of their bit so that when it is actually incorporated it will not be their widget that has spoiled the party. Aero guys have been known to pad their drag estimates and reduce their lift estimates for similar reasons. I for one will only believe the weight of an aircraft once it has been built and weighed.
The Hawk was one design that performed better than estimated and I believe this was (at least in part) because of this. It was also the first Kingston fixed price contract with penalty clauses for failing to meet spec which probably increased the motivation mentioned above.
Of course there are some who use the opposite technique. They say it will weigh less than they know it will, will cost less than they know it will and rely on the customer not feeling the difference is big enough to cause cancellation.....but perhaps that sort of thing is more related to company culture than engineering?