This question gives written premiums and earned premiums by calendar year, and net loss ratios on an occurrence basis, from which you need to calculate underwriting year loss ratios. The answer in the revision booklet makes the assumption that 2/5ths of business is earned in the year it is written, and the rest is earned the next year. Given business started in the first year for which we have data though, can't you calculate how much of the earned premium for each year relates to which underwriting year directly, by assuming all premium earned in the first calendar year related to business written in the first calendar year, and policies are no longer than annual?
Yes you can, but that's where the 2/5ths comes from. 10/25 = 2/5ths. the remaining 15 must be earned in 1999, so (35-15) / 50 = 2/5ths etc. I suppose you could argue it's not an assumption.
