Hi, Does anyone know of a neat (regression??) solution to the following problem?: GLR = GC / GP RLR = RC / RP = ( GC x CR% ) / ( GP x CP% ) NLR = NC / NP = ( GC - RC ) / ( GP - RP ) Regression variables: NLR = y GC = x1 CR% = x2 CP% = x3 Where: GLR = gross loss ratio GC = gross claims GP = gross premium (arbitrarily set to a fixed value of 100 for the purpose of simplifying the problem) RLR = reinsurance (ceded) loss ratio RC = reinsurance (ceded) claims recoveries RP = reinsurance (ceded) premiums CR% = ceded recoveries % (of gross claims) CP% = ceded premiums % (of gross premiums) NLR = net loss ratio NC = net claims NP = net premiums