Ouch! very fewer marks for this odd question. Understanding given information:- Let Y= aggregate loss cost, and \( x_i \)= loss from \(i^{th}\) case in the period of coverage Note: a case could be an individual loss or a loss from individual event, not given in the question. That won't alter the answer anyway. So, \(Y=\Sigma_i max(x_i-1000000,0)\) Now, table is given for \(LEV_Y(Aggregate~Limit) \) Note: here LEV is based on absolute numbers rather than relative to EML/PML. there are 2 implicit assumptions in the given answer that I could think of. 1) \( x_i <= 2000000 \) for any i. Otherwise, that loss will be uncovered before any reinstatement occur. Or, maybe it is possible to allow for a loss in the layer before and after reinstatement in part. 2) there is no payable reinstatement allowed. Otherwise, calculations are not straightforward. Answers: a) from unlimited free reinstatements - there is no aggregate limit from AAD - first 1000000 will be reduced that would otherwise be counted as the final loss So, LEV(Unlimited)-LEV(1000000) or E(Y)-LEV(1000000)=1289000-350000=939000 b) from 1 free reinstatement - the aggregate limit is twice of the given loss layer. So, LEV(2000000)=525000 c) from AAD - first 1000000 will be reduced that would otherwise be counted as the final loss would be counted as the loss is answered in 'b)' So, LEV(2000000)-LEV(1000000)=525000-350000=175000 PS: I don't know how the first alternative to 'c)' is calculated. maybe, they took Deductible as aggregate excess besides case excess.