Errata for the 2019 Exam 8 Cookbook
I will post any corrections to the Exam 8 Cookbook here. Please check back occasionally.
If you see an error in the Exam 8 Cookbook, please send an email to exam8@RisingFellow.com. Thanks!
GLM - Comparing Models: F-Test (9/22/19): In Step 3, the denominator DoF should be the degrees of freedom from the Big model. The calculations are correct, but the formula subscript still incorrectly said small. Below is the correct formula that matches the CAS errata for the paper:
Summary of Formulas Correction (9/22/19): I corrected the summary of formulas. It still incorrectly had the wrong denominator DoF for the F-Test. The denominator DoF should be based on the BIG model. I fixed it so that it’s consistent with the GLM errata now and you can download it the correct summary of formulas here.
Fisher - LDD Premium (9/22/19): In the sub-bullet on pg. 108, the sentence should read: "...whereas a $1M limit LDD with a $100k deductible covers the layer $900k x $100k". The word aggregate shouldn’t be there since this is discussing the policy limit at the per-occurrence level.
Clark - Casualty Excess of Loss: Exposure Rating (9/9/19): [Clarification Note pg. 137] Reasoning for the .077 in the “ALAE Included with Loss Example:”
If you look at the bottom of the previous page (pg. 136), the %ALAE in the formula is the ALAE as % of loss capped at the Policy Lim. Since the ELR is 65% and the 5% ALAE is the % of premium, the %ALAE for the formula can be calculated as 5% / 65% = 7.7%.
Mahler - Credibility-Weighted Predictions X_est (8/21/19): The PDF cut off the mu in the the formula under the section “Credibility Weight of the Most Recent Data and Previous Estimate”. Below is the correct formula with mu at the end:
ISO (2014) - Experience Rating (8/8/19): The 2019 CAS study kit has an updated ISO paper. The only significant changes are new numbers in some of the tables. I fixed the Experience Rating recipe to reflect the new numbers. I updated the digital version of the Cookbook with this change (and the updated F-Test recipe), but you can also download it directly below:
NCCI 1 - Experience Mod (7/29/19): The experience mod in step 6 should be rounded to 2 decimal places according to the rating manual. The experience mod should change to 1.47. This changes the final standard premium to 119,055.
GLM - Comparing Models: F-Test (7/29/19): The CAS posted an updated errata for the GLM paper (link to online text references). In it, they modify the F statistic formula and differentiate between the scaled and unscaled deviance. Make sure to review the errata for the paper. I fixed the recipe with the new errata. You can download it here: