If Carlsberg provided actuarial advice, we're sure it would be the best. One thing they would be bound to tell us is that noone on Earth knows what the future holds. That actuaries sometimes pretend they have some special talent in this area is bizarre.
In particular, we don't know when, in which direction, how far or for how long, financial conditions will change. Demographic changes tend to be rather smoother so there is reasonably more room for good guessing about the future than for the financial elements.
Even so, we need to do our best, trying to make assessments over what can be very long periods, such as 50 years or even longer. Basing assumptions for the very long-term upon short-term statistics seems extremely odd. On the other hand, for very long periods, there may have been major structural shifts, making it difficult to justify using few very long periods for analysis. Accordingly, we think it is entirely appropriate to compromise with as many decent time chunks, lasting say 15 years, as we can find, even if they overlap.
Further, using scalars to represent many future possibilities provides stakeholders with no idea of the variations; the certainty offered is non-existent. This is explored further here.