CAS Exam Errata

We always hated seeing errors in study material when taking exams, so we check everything as much as possible. If you do see a mistake, check out the up-to-date errata below for any corrections.

Accidents Happen ...

If you spot an error in the study material, please send us a message so that we can correct the error and document it here for others.

Errata by Exam

Study Guide & Review Videos

WM Ch. 5 – (2/8/24) – On page 51, at the end of the premium audit example, the calculation has a typo. It should say “At 12/31/13: All policies have been audited. The premium is $500,000(1.08)(12) = $6,480,000.”

Friedland Ch. 5 – (2/25/24) – On page 347, the raw payroll is shifted down by one starting in 2005. The 2005 payroll should be 350,000, 2006 payroll should be 790,000, etc. The video has the same error. We will repost the video ASAP.

Cookbook

Reserving – ULAE: Classical Technique (3/22/24) – The example calculation in the Discussion section is incorrect. Case Outstanding should be multiplied by 60% because the ULAE for closing is outstanding.

It should be (Case Outstanding)*(60%)*(ULAE Ratio) + (Pure IBNR)*(100%)*(ULAE Ratio) = (301,240)(0.60)(0.053) + (99,712)(1.00)(0.053) = 14,740.

Ratemaking – Loss Aggregation (2/20/24) – Part b of the problem should state “Build an accident year reported loss cumulative development triangle,” not a paid loss triangle.

Reserving – Frequency-Severity Technique #2 (1/2/24) – In Step 4, the formula for Estimated Frequency(prior year) has an error in it. In the numerator, the Selected Frequency should be for the latest year.

Ratemaking – Loss Ratio Method (1/2/24) – The description of Step 7 is incorrect and refers to the pure premium method. Instead it should state: “Calculate the indicated rate change using the loss ratio indicated rate change formula.”

Past CAS Problems

Spring 2018 Q #8 – (3/28/24) – The CDFs calculated should have included the given tail factor (1.067). This was an oversight. We fixed the solution and re-uploaded the corrected version.

Spring 2018 Makeup Q #1a – Policy E should not be included because its policy inception is after 9/15/15.

Spring 2018 Makeup Q #10 – The Limited Average Severities for the 50-100k and 100-250k layers were incorrect. They should be unconditional so the denominator should be the total number of claims used for the calculation. The problem is updated and re-posted.

Spring 2018 Makeup Q #11 – For part a, the problem asks for the PLR. The original solution calculated the variable expense ratio, but failed to calculate the final PLR. The problem is updated and re-posted.

Spring 2018 Q #18 – The CDFs being applied to calculate the final ultimate severity for 2017 are shifted by one. The CDFs for 2014 – 2016 should be 1.000, 1.042, and 1.130, respectively. This changes the final ultimate severity for 2017 to $3,243.

Study Guide & Review Videos

CFR Ch. 7 – (3/8/24) – In the solution to part c. on page 275, the UEPR should only include the first two columns from the U&IE. Hence, UEPR = 2,000 + 500 = 2,500. This changes the final answer to part c. to $8,450,000. Although Odomirok sums all four columns when showing the UEPR on the U&IE, the last columns get subtracted out at the very bottom of the U&IE, which means the final UEPR that makes it way to the balance sheet is just the sum of the first two columns. I will be re-doing this section of the Ch. 7 video to make that more clear. See 2014 #14 for another good example of calculating UEPR from the U&IE.

CFR Ch. 19 – (3/14/24) -On page 562, the description of the RBC charge for an investment affiliate should simply say “the RBC charge for an investment in an investment affiliate is 0.225 times the carrying value of the common and preferred stock, adjusted for the owernship in the investment affiliate.” In our opinion, there is an error in this section of Odomirok. In 2016, the NAIC moved from the “look through” approach (described in the second paragraph of this section in Odomirok) to the “22.5% of the carrying value of common and preferred stocks.”

NAIC Preamble – (3/18/24) – On page 795, the bullet around precision should say “The degree of precision matters. If a judgment item can be precisely estimated, the materiality threshold should be lower.” On page 798, the solution to 3b should say “MORE precision leads to a lower materiality threshold.”

Porter 6 – (3/25/24) – On pages 95, part a. should ask for the “minimum requirement for this insurer.” On page 96, the solution should say “the minimum requirement is the sum…”

Flashcards

NAIC Preamble – (3/18/24) – On card 3 (page 6 of standard PDF, page 2 of printable PDF), the last sentence should say “…significantly different outcomes in regulatory requirements.” On card 4 (page 8 of standard PDF, page 4 of printable PDF), it should say “Lower materiality thresholds should be used when there is more precision in estimating errors.”

Study Guide & Review Videos

Friedland (12/7/23) – There is a typo on page 142. The title of the third table should be “Abs. Diff. — Reptd. Age-to-Age Factors.”

Taylor & McGuire (1/21/24) – On the Over-Dispersed Poisson slide of the video, the factorial after “(y/phi)” was missing in the first few lines. The video has been corrected and reposted.

Venter Factors (1/25/24)  The conclusion shown in the Venter Factors Stability Tests workbook (located under “Extras for Download -> Paper Exhibits & Other Examples” in the Online Course)  for Stability Test #2 is not valid. Since the age-to-age factors do not appear to rotate around some fixed level, we should give more weight to the most recent age-to-age factors when making a final selection. The workbook has been corrected and reposted.

Venter Factors (1/25/24)  In MP #1 part c., the language used in the problem prompt is misleading. The word “improving” should be replaced with “dealing with.” The point is that using a weighted average of age-to-age factors does not mean the instability improves. It’s just a way of handling it.

Verrall (2/16/24) – On page 739, the first disadvantage of the ODP model should just say, “It requires the column sums of incremental values to be positive to ensure a positive variance.”

Venter Factors (2/26/24) – On page 343, in the second to last bullet, it should say “left with four f(d) parameters and h = 1.”

Cookbook

Shapland – Expected Incremental Losses from GLM (2/29/24) – In the Excel version, the second formula in the discussion section should show:

nu_1,d = c + sum(betas)

Shapland (1/25/24) – In the Excel version of the “Negative Incremental Values” recipe, the solution to part a. is pointing to the incorrect set of incremental losses. The adjusted log-link triangle should be pointing to the incremental losses in cells D8:F10. The workbook has been corrected and reposted.

Past CAS Problems

Venter Factors 2016 #6 – The second 48-60 LDF of 1.027 shown in the triangle should not be there.

Problem Bank

Siewert (2/7/24) – In the first line, it should say that the deductible is 100k (not 500k).

Taylor & McGuire (3/13/24)– In the solution to RF #7a, the condition of “X_k,j” should not be there. It should just say “Y_k,j ~ …”

Marshall (3/16/24)– In the solution to RF #2b, the WC CoV chart should be 24%, 16%, 10%, and 6% so that it is NOT linear.

Flashcards

Taylor & McGuire (2/2/24) – On card 11 (page 22 of standard PDF, page 8 of printable PDF), the condition of “X_k,j” should not be there.

High-Level Summaries

Siewert (3/23/24) – “Deductible Loss Charge” should instead be “Per Occurrence Charge” (chi).

Taylor & McGuire (2/2/24) – On page 50, under the “Cross-Classified Model” Assumptions, the condition of “X_k,j” should not be there.

Exam 8 Cookbook

Clark – Property Per Risk: Exposure Rating (9/29/23) – In step 2, the example calculation should show a subscript of 100-250 for the Expected Reinsurer Loss.

Practice Exam

Problem 13 (10/7/23) – The 2023 version of the study kit no longer has some of the tables, so they would need to be given in the problem itself.

Practice Problem Bank

Clark – 13 (10/7/23) – Excess loss factors by definition can’t be greater than 1. Some of the given a and b parameters result in ELFs greater than 1. This is a mistake in the parameter values given. It doesn’t change the approach to solving the problem though. I will fix the values given in the problem for next sitting.

NCCI Circular – 1 (10/7/23) – The 2023 version of the study kit no longer has the two tables to look up the expected claim count group (ECG) based on the number of expected claims and subtable based on the per-occurrence excess ratio (k).

On an exam question, you would need to be given something similar to look up the correct ECG column and subtable. More likely they would just give you an excerpt from the appropriate column and subtable and skip the lookup.

Below are screenshots of those lookup tables so that you can see what they look like:

Excess Ratio Subtable Lookup

Expected Claim Count Group Lookup

Bernegger – 2 (10/12/23) – In the part b solution, the first and second derivatives of G(x) are incorrect. I updated the derivative calculations and uploaded the corrected version. The conclusion of part b that G(x) is a valid exposure curve is unchanged.

There is an alternate solution by using the formula:

E[x] = E[X/MPL] = 1 / G'(0)

This results in a different answer (~112,000). The reason for the inconsistency is that the values we have for gross premium and expected ceded loss aren’t internally consistent. I need to rework this problem a little with the values given. You should get the same MPL either way.

Fisher – 11 (9/29/23) – In the part a solution, the entry ratio label should say XS of 1.8 (not XS of 1.15). The calculation is correct.

Grossi – 1 (7/8/23) – Part a should also say that the AAL would increase with the higher hurricane loss. The files have been updated.

Grossi – 9 (7/8/23) – Similar to the example in Grossi (Table 5.1), all values in the table should be “per $1,000” of property value. The files have been updated.

Study Guide & Review Videos

The first five errata were posted before we uploaded the revised study guide due to the changes in Exam 9 Content Outline. If you downloaded the study guide after December 6, you can ignore the first five errors below since they were corrected in the revised guide.

Clark (11/8/23) – There is a typo on page 8. In the paragraph below the table, the second line should say “1.03^5 = 1.159,” not 1.086.

Clark (11/12/23) – There is a typo on page 34. In the fourth bullet of the Swing Plan example, the minimum premium should be 10%, not 15%.

Clark (11/29/23) – There is a typo on page 29. In the General Liability exposure rating example, it should say that “an actuary is pricing a 7.5M XS 2.5M excess of loss treaty.” We will correct the video as well.

Mildenhall Ch. 4 (12/1/23) – On page 123, an extra row needs to be added at the bottom of the table in the example. When p = 0.9, VaR = 12, TVaR = 25, and CTE = 25. When p > 0.90, VaR = TVaR = CTE = 25.

Clark (12/5/23) – A few items were included in the Clark outline that are not included in the most recent version of the paper. The following items have been revised:

  • Page 5 – Discussion under step 5
  • Page 8 – Discussion under step 4
  • Page 41 – Discussion under pricing cat covers
  • Pages 45 – 46 – Discussion around ROE and allocated surplus have been removed

The following errata were posted after we uploaded the revised study guide on December 6.

Clark (12/12/23) – On page 18, the Trended, Developed Treaty Losses should be $1,071,280.

Mildenhall Ch. 4 (12/22/23) – In the crossed pairing example on page 113, we should only be crossing the top 4 values of X_1 and X_2. Thus, the pairs should be (52, 8), (36, 12), (18, 23), and (10, 40). These pairs produce X values of 60, 48, 41, and 50. Then, VaR_0.70(X) = 41. See the forum discussion on this for more details.

Clark (1/16/24) – On page 36, the Poisson probability formula has an error in it. The formula should be lambda^(n) * exp(-lambda) / (n!).

Grossi Ch. 6 (1/17/24) – On page 77, the final amount ceded to the XOL treaty should be 0.50 x MIN(6.8 – 2.0, 4) = 2.0. The final net loss to the insurer is $6.8M – $2.0M = $4.8M.

Mildenhall Ch. 9 (1/17/24) – On page 158, at the top of the page, the “Net X2” value for the second row in the table should be 1. On page 161, under the second step, the alpha shown should be 100, not 50. The calculated VaR shown in step 2 is correct.

Brehm Ch. 5 (1/17/24) – On page 330, under the “Supply and demand” section, the last sentence in the first bullet should say “better performing business,” not better performing firms.

Mildenhall Ch. 10 (1/20/24) – On page 167, the “and h” should be removed from the last bullet. On page 169, under the big Example in the middle of the page, the second bullet should be h(p) = h(0.05) = 1 – g(0.95) = 0.0203. Since this is the bid price, 1_U>0.95 means that the probability of an investment payout of $1 is 0.05. Hence, p = 0.05 for this bullet since h(p) is the value of a bond with a Bernoulli payout have probability p of full payment.

Clark (1/21/24) – On page 18, there is a typo at the end of the fourth step. The words “expected treaty loss ratio” should say “expected treaty loss cost.”

Mildenhall Ch. 4 (1/25/24) – On page 113, the crossed pairing in the second step is incorrect. Since there are 10 values and p = 0.70, we only care about the top 4 values (since the fourth largest loss is VaR_0.70). We should pair the largest X1 value (52) with the fourth largest X2 value (8), the second largest X1 value (36) with the third largest X2 value (12). This produces a VaR_0.70(X) = 41. You can find a more detailed discussion of this example on the course forum. We corrected this example in the video.

Grossi Ch. 6 (1/25/24) – On page 78, the guide implies that we can re-calculate OEP($100M) by subtracting the three earthquake events from the original OEP of 0.557%. However, this is only an approximation. In reality, we would need to do the same OEP math covered in Grossi Ch. 2. You can find a more detailed discussion of how to calculate OEP($100M) after removing the EQ events in the course forum.

Clark (2/1/24) – At the top of page 38, the second term in the variance equation should say “Var(Freq) x [E(Sev.)]^2.”

Grossi Ch. 6 (2/1/24) – On page 77, in the first bullet, the final loss should be $2.8M.

Mildenhall Ch. 4 (2/1/24) – There is an error in the textbook that made its way into the guide. On page 108, the definition of the n-year occurrence PML should say “at most 1/n instead of “at least” 1/n.

Mildenhall Ch. 10 (2/10/24) – Minor typos. On page 207, for j = 7, S should be 0. On page 209, the first step should show 0.313^0.5 = 0.559. On page 213, the denominator of bullet 2 should be 0.441.

Mildenhall Ch. 5 (2/10/24) – Clarification for the definition of comonotonic additive on the top of page 128. If X and Y are comonotonic AND rho(X+Y) = rho(X) + rho(Y), then rho is comonotonic additive (COMON).

Mildenhall Ch. 5 (2/20/24) – On page 130, the fourth bullet under Dual Utility Theory should say “Dual utility theory is linear in outcomes based on distorted probabilities.”

Mildenhall Ch. 9 (3/5/24) – On page 160, the first bullet in the takeaway section should say “if two units are the same size, the riskier unit will have a HIGHER gross premium and LOWER gross loss ratio.”

The following errata impact the new version of the guide uploaded on 3/12/24. All prior errata were fixed in the 3/12/24 version.

Mildenhall Ch. 4 (3/15/24) – On page 118, the definition shown for CTE is for the Lower CTE. The definition of the upper CTE is E(X|X >= q^+(p)). The example shown further down in the page is calculating the upper CTE. We have reposted the video to make this clear.

Mildenhall Ch. 4 (3/18/24) – On page 118, the Lower CTE calculation should say (4+8+12+25)/4 = 12.25. The Upper CTE calculation should say (8+12+25)/3=15.

Clark (3/27/24) – On page 33, under the formula for the exposure factor for drop-down coverage, it should say that “the first two terms in the numerator are capped at UL + PL and the last two terms in the numerator are capped at PL.”

Flashcards

Mildenhall Ch. 4 (1/17/24) – On card 16 (page 32 of standard PDF, page 12 of printable PDF), the general relationship is incorrect. It should say “Var <= CTE <= WCE <= TVaR.” The flashcard PDFs and Anki deck have been reposted with the corrected card.

Mildenhall Ch. 10 (1/20/24) – On card 15 (page 30 of standard PDF, page 24 of printable PDF), the “and h” should be removed from the fourth bullet. The flashcard PDFs and Anki deck have been reposted with the corrected card.

Clark (2/1/24) – On card 35 (page 70 of standard PDF, page 10 of printable PDF), the second term in the variance equation should say “Var(Freq) x [E(Sev.)]^2.”

Bernegger (2/2/24) – On card 7 (page 14 of standard PDF, page 6 of printable PDF), the first term in the second equation should be G((AP + Limit)/MPL), not G(AP + Limit)/MPL.

Mildenhall Ch. 10 (2/5/24) – On card 15 (page 30 of standard PDF, page 10 of printable PDF), the final bullet point is incomplete. It should say “discount uncertain assets. Thus, h(p) <= p and g(s) >= s.”

Mildenhall Ch. 5 (2/10/24) – On card 10 (page 20 of standard PDF, page 8 of printable PDF), the definition of comonotonic additive needs to be clarified. It should say “If X and Y are comonotonic AND rho(X+Y) = rho(X) + rho(Y), then rho is comonotonic additive.”

Mildenhall Ch. 5 (2/20/24) – On card 16 (page 32 of standard PDF, page 12 of printable PDF), the second bullet should not have the word “not.” On card 18 (page 34 of standard PDF, page 12 of printable PDF), the fourth bullet should say “Dual utility theory is linear in outcomes based on distorted probabilities.”

Mildenhall Ch. 4 (3/15/24) – On card 15 (page 30 of standard PDF, page 10 of printable PDF), we added the upper CTE definition.

Clark (3/27/24) – On card 28 (page 56 of standard PDF, page 20 of printable PDF), it should say  that “the first two terms in the numerator are capped at UL + PL and the last two terms in the numerator are capped at PL.”

Problem Pack

Sahasrabuddhe 4 (3/25/23) – In part b, the numerator in the formula should show (LEV[ X | Theta_AY,ult ] / LEV[ B | Theta_AY,ult ] ). This was a correction in the Sahasrabuddhe paper errata (equation 3.10) and I didn’t update this formula in the problem.

The actual solution is correct though.

Bernegger 2 (1/26/23) – Minor errors. For conditions 3 and 4, it should state that g > 1 for this MBBEFD curve. Also, the quantity “1+(g-1)*x” in the denominator should not be squared in the G'(x) formula.

Mildenhall Part II – 3 (1/2/24) – The Cookbook formula for WACC had an error in it. This problem is updated to fix the error in the reinsurance cost formula.

Mildenhall Part I – 1 (1/17/24) – The original Mildenhall Part 1 problem bank had an error in the CV calculations for #1. The problem bank has since been reposted with the errors corrected.

Mildenhall Part I – 8 (2/1/24) – The solution, while correctly using the probability of 0.25, incorrectly notes 0.1 as the probability used in part a. Problem pack has been updated.

Mildenhall Part I – 12 (2/1/24) – Some grammar mistakes caught and updated. Solution did not change, but question and solution now read better.

Mildenhall Part III – 19b (2/1/24) – Because the calculations in the table are based on unlimited losses, the headers should not read X(a), etc. The headers have been updated and the file posted.

Mildenhall Part I – 11 (2/1/24) – The 0.75-VaR for X should be 90, not 75, and for X’ 140, not 125. The remainder of the solution is consistent.

Mildenhall Part I – 12 (2/1/24) – The TVaR for X + Y should be the sumproduct of the probabilities and loss amounts for the four losses above the threshold, not the simple average. The file has been updated and reposted.

Mildenhall Part I – 4 (2/1/24) – Definition of crossed pairing has been updated and the resulting solution has changed. The file has been reposted.

Mildenhall Part I – 8 (2/19/24) – Part a. should be asking for the assets (as shown in the solution), not the capital. The file has been reposted.

Cookbook

Mildenhall Part 3 – Allocated CCoC – Implied Target Loss Ratios (3/22/24) – I modified steps 2 and 3 a little bit to put the right formulas so that it’s easier to follow. Nothing changed with the solution, but the original version made it seem like WACC = risk discount rate (delta), which isn’t correct. Instead, WACC = Cost of capital (iota).

Download the updated Excel version to see the clean version.

Clark – Recursive Aggregate Generalized (2/27/24)- The formula for b and Pr(n) should show alphas instead of ‘a’ in the formula.

Portfolio Constant Cost of Capital Pricing (2/22/24)- The final formula should show:

Cost of capital=Target return on capital × Amount of capital

Insurer Insolvency and Liquidation (2/22/24) – In step 1, it should say “Liabilities > Assets, so the insurer is bankrupt.”

Allocated CCoC – Implied Target Loss Ratios (2/19/24) – Minor typo. In the PDF, the calculation shown in step 1 should show “= 9,350 x 37%”

Bernoulli Layers & Distortions / Spectral Risk Measures (2/17/24) – For the Wang Transform Excel formulation, the sign was wrong and should read “+ lambda”:

g(s) = NORM.S.DIST( NORM.S.INV( s ) + 𝜆 , TRUE )

One-Period DCF Model (2/17/24) – The prompt should state “net premium (excluding expenses)” not excluding taxes.

Weighted Average Cost of Capital (WACC) (1/18/24) – The Cookbook formula for WACC had an error in it. The numerator for the reinsurance cost should be: Ceded Premium x (1 + risk-free rate) – E[Recoveries]

This is because the loss of investment income from paying the ceded premium at t = 0 is an added cost.

Mildenhall Part III – Euler Gradient (2/1/24) – The problem listed the incorrect coefficients. It should now read 𝜌(𝑃,𝑅,𝑎)=√((0.36𝑃)^2+(0.34𝑅)^2+(0.08𝑎)^2 ) to match the solution.

Mildenhall Part II – Classical PCPs (2/1/24) – In the discussion section, the second formula for semivariance should note E(X – E[X])^2 as the formula for variance.

Past CAS Problems

CAS 2011 #9 (Exam 8 Bernegger) – The numerator of the G(x) function should be ln(a+b^x) – ln(1+a).

CAS 2014 #24 (Exam 8 Grossi) – The solution incorrectly used the payback pure premium from part a to calculate the reinstatement premium. It should use the actual premium given in part b ($1.2M).

High-Level Summaries

Midenhall Part II – Ch. 8 (2/19/24) – On page 30, it incorrectly states that IFRS loss reserves are undiscounted. Should say IFRS loss reserves are discounted including a risk adjustment.

Mildenhall Part I – Ch. 5 (2/20/24) – On page 25, the second row of the table under Utility Theory should not have the word “not.” The fourth row of the table under Dual Utility Theory should say “Linear in outcomes based on distorted probabilities.”

Formula Summary

Portfolio Constant Cost of Capital Pricing (2/22/24)- The final formula should show:

Cost of capital=Target return on capital × Amount of capital

Rising Fellow ensures you have all of the materials you need to confidently sit your next CAS exam.

Geoff Werner joining Rising Fellow for Spring 2024 Exam 5 Online Course »