Hi vikky
I suspect you're not alone. This is certainly not the clearest thing in the world 
When we're considering a GPV, we have a reserve of the form:
reserve = value (benefits) + value (expenses) - value (premiums)
where the premiums are the actual premiums (ie the office premium).
If we consider the actual premiums for a WP policy, we'd expect them to be higher than for an otherwise similar without-profits policy. The WP policy premiums include some "bonus loading", ie the assumptions used in calculating the premiums are such that the premiums should be more than enough to cover the sum assured and expenses, and this "extra" can be distributed as bonus.
Looking back at the reserve formula, the value of these future premiums (with their bonus loadings) is subtracted from the reserve, so making it smaller. I think the Core Reading would describe this as the method capitalising the loadings in future premiums.
We therefore need to be careful to also include enough of an allowance for future bonuses in the GPV reserve formula. So, the "value(benefits)" term should include not just the sum assured and past bonuses, but also a prudent allowance for future bonuses consistent with PRE and the loadings in the premiums. Determining what is an appropriate allowance is not straightforward.
The net premium valuation (NPV) avoids much of this. Remember that a net premium reserve has the form:
reserve = value (benefits) - value ( net premiums)
where the net premiums are calculated using the same assumptions as the reserve. We no longer have the consideration of the GPV method about the relationship between the reserving basis and the premium basis. In the NPV, we typically allow for future bonuses by using a lower future valuation rate of interest.
I hope this helps a little.
Best wishes
Lynn
