exam answers

[root4unc]

When will the CAS release the exam and post the solutions? I know that the multiple choice answers are usually released at the same time as the exam.

[Wesley_Willis]

When will the CAS release the exam and post the solutions? I know that the multiple choice answers are usually released at the same time as the exam.

WTF was with the dearth of accounting? I guess Mr. All 10 led us astray with his projected threat of even more accounting than in previous exams at the seminar. Thanks dude...

[Utanapishtim]

But the new accounting note was on there, and as predicted, it was about 'paradigms'!

[ed999]

what ARE the four paradigms? were they talking about GAAP, SAP and TAX Rules???

[Avi]

I read Blanchard > 2.5 months ago and forgot to recram it :(

I think I just listed GAAP, SAP, Lloyds, and something else.

I consider that question a 0 in my internal count :(

[Utanapishtim]

I thought he meant the paradigms for the 4 different users -- Investors, Regulators, Tax Regime, and Management.

If not, I'll have to count that one out...

[the yellow dart]

That's one I got: GAP, SAAP, Tax, and Internal.

Was I missing something obvious with the Mac question? The Paid LDF threw me off.........

I'd like to think I have around 50-55 points, but if I screwed up the multiple choice section again........

[Avi]

The direct/indirect XS methods. Were both 1.2 M?

Was R_L 100000/166667 or .8?

[Argh!]

I thought he meant the paradigms for the 4 different users -- Investors, Regulators, Tax Regime, and Management.

If not, I'll have to count that one out...


That's how I answered the question as well. Hopefully that's it!

[Wesley_Willis]

But the new accounting note was on there, and as predicted, it was about 'paradigms'!

Hmm. I expected the Blanchard question to be about the six requirements of accounting (Understandable, Reliable, Relevant, etc.)

I still can't figure out what the four paradigms should be after scanning through the All 10 manual again. I'm sure it's explicitly stated in the article somewhere, but that would require me to read the article...

:horse: :horse: :horse:

[Argh!]

Anyone care to share their multiple choice answers?

[ramanujan]

The direct/indirect XS methods. Were both 1.2 M?

Was R_L 100000/166667 or .8?

I used change in relativity = R_L = .8/(100000/166667). My answer was definitely different for direct.

[Avi]

For indirect I got 1.2 M

For direct, don't we need to solve:

LDF = LDF_L*R_L + XSLDF_L*(1-R_L).

Isn't ΔR_L used for LDF_L=ΔR_L*LDF?

If my logic is correct, then I wasn't sure whether to use .8 (8M/10M) or 100000/166667 as R_L. Using the former, the XSLDF is 3.00. Using the latter the XSLDF = 3.4. Using 3.4 gives an expected IBNR of 2.4*500000 or 1.2M

What I used to decide was when they asked for the credibility method, although i totally blanked on the LR method, isn't the credibility method a weighting of LR and projection? If so, since the Direct vs. Indirect was not specified, doesnt that imply that they are the same?

[Avi]

Anyone care to share their multiple choice answers?

C
E
C
E
D
A
skip
E
skip
skip
B
E
C
A
C
B
C
E
B

[lee]

For indirect I got 1.2 M

For direct, don't we need to solve:

LDF = LDF_L*R_L + XSLDF_L*(1-R_L).

Isn't ΔR_L used for LDF_L=ΔR_L*LDF?

If my logic is correct, then I wasn't sure whether to use .8 (8M/10M) or 100000/166667 as R_L. Using the former, the XSLDF is 3.00. Using the latter the XSLDF = 3.4. Using 3.4 gives an expected IBNR of 2.4*500000 or 1.2M

What I used to decide was when they asked for the credibility method, although i totally blanked on the LR method, isn't the credibility method a weighting of LR and projection? If so, since the Direct vs. Indirect was not specified, doesnt that imply that they are the same?

I get the XSLDF = 3.4 using the same way as yours.
I made the B-F method is credibility weighting of incurred loss amount and LR ratio method.

[Avi]

I agree on WHAT the cred weighting is. I just blanked as to how to do the LR method :(

How DO you do it?

I'm sure I'll kick myself when I see it :(

[lee]

I agree on WHAT the cred weighting is. I just blanked as to how to do the LR method :(

How DO you do it?

I'm sure I'll kick myself when I see it :(

Expected Loss = earned prem* ELR*X, which means the expected loss fall into the layer above deductibles.

So Expected Loss= expected unlimited loss - expected loss below the deductible.

Is this logic right? But I am not sure which LDF I should use to credibility weighting. I used LDF (the unlimited one).

[Avi]

Kick
Kick
Kick
KIck
Kick
Kick

[islas_del_maiz]

Here I thought the relativity they gave (1.66667) was for reported as of 12 mo or sth. So I did a Pinto-Gogol step:

excess LDF=(10-8)/(10/unlim LDF - 8/lim LDF)

I think this makes sense but not sure if it's intended.

For indirect I got 1.2 M

For direct, don't we need to solve:

LDF = LDF_L*R_L + XSLDF_L*(1-R_L).

Isn't ΔR_L used for LDF_L=ΔR_L*LDF?

If my logic is correct, then I wasn't sure whether to use .8 (8M/10M) or 100000/166667 as R_L. Using the former, the XSLDF is 3.00. Using the latter the XSLDF = 3.4. Using 3.4 gives an expected IBNR of 2.4*500000 or 1.2M

What I used to decide was when they asked for the credibility method, although i totally blanked on the LR method, isn't the credibility method a weighting of LR and projection? If so, since the Direct vs. Indirect was not specified, doesnt that imply that they are the same?

I get the XSLDF = 3.4 using the same way as yours.
I made the B-F method is credibility weighting of incurred loss amount and LR ratio method.

[Prawnz Perdinand]

Anyone care to share their multiple choice answers?

C
E
C
E
D
A
skip
E
skip
skip
B
E
C
A
C
B
C
E
B
I think you got the "skips" right.

[Avi]

Thanks :)

[Wesley_Willis]

I thought I had done well on the MC questions, but upon further reflection and discussion with my peers, I managed to botch plenty of them.

[carenthir_dm]

C
D
C
E
D
A
C
E
Skip
Skip
B
Skip
C
A
Skip
B
C
E
B

[Avi]

So, will you post your choices and perr consuses? :)

[Father of two]

C
C
A
E
E
A
SKIP
E
SKIP
D
B
D
C
A
C
B
C
E
SKIP

I should have guessed on some of my skips.

[lori darling]

wow, i got a whole lot more "d"s than that for the first like ten answers.

[idomoneus]

1 C
2 C
3 C
4 D
5 C
6 A
7 skip
8 E
9 skip
10 D
11 B
12 E
13 C
14 A
15 B
16 B
17 B
18 B
19 E
20 E

For some of the questions there is starting to be a consensus, but we need more data points...

[lee]

1 C
2 C
3 C
4 D
5 C
6 A
7 skip
8 E
9 skip
10 D
11 B
12 E
13 C
14 A
15 B
16 B
17 B
18 B
19 E
20 E

For some of the questions there is starting to be a consensus, but we need more data points...
There is no #20

[KindGrind]

Hey guys, I got all the data from you guys, so there's 6 ppl from which I got answers. So feel free to send answers! I'll give you guys feedbacks about how I think you guys scored!

-Crystina

[idomoneus]

I had

18 E
19 E

[KindGrind]

yeah that's what I guessed from the answer key!
If you want to know how much you got at the moment (kinda expected value) just ask and I'll PM you guys

[Argh!]

Did anyone write down their numerical answers to any of the essay questions? I'll share the ones that I wrote down.

21.) 122,908
23.) 6,723,406
25.) 972,040
29.) a.) .3, .05, .05
b.) 1301
c.) redundant by 2,222,000
30.) 1,491,066
31.) 2002 - 16,461
2003 - 12,234
33.) a.) 2001, 3M
b.) 1/3
c.) 6.4M
34.) 5115
35.) a.) 754
b.) 770
c.) 755
36.) 17,965
39.) A-506
B-667
C-500
42.) a.) 243, 243
b.) 186, 300
43.) a.) .2024
b.) .4789
45.) d.) .125
47.) 1,875,000
49.) a.) 2100, .646
b.) 2168.4, .657

[ramanujan]

I finally got my test back. Here are my mc answers.

1 C
2 C
3 C
4 D
5 D
6 A
7 B
8 E
9 D
10 skip
11 B
12 E
13 C
14 A
15 C
16 B
17 C
18 E
19 E

[ed999]

i don't see how # 2 is C.
Here's how i did it...

I just took the % unreported. Took the value of the incremental % unreported and summed it to get .58...

I guess one call also approach it as age to age - 1 = 1.6-1 = .6 which is also close to .58 and choice E.

what do u think??

[ramanujan]

i don't see how # 2 is C.
Here's how i did it...

I just took the % unreported. Took the value of the incremental % unreported and summed it to get .58...

I guess one call also approach it as age to age - 1 = 1.6-1 = .6 which is also close to .58 and choice E.

what do u think??

I think the question is asking - what % of IBNR will emerge in during period 12-24 as a percentage of IBNR at age 12?

% unreported at 12 = .584
% unreported at 24 = .334

(.584 - .334)/.584 = .428

which is answer choice C.

[ramanujan]

My answers for 7 and 9 seem to be different from others. Here is explanation for my answers.

7. If you take the premium and losses @18 months to calculate PDLD ratio, you will get answer choice C. If you take premium and losses @27 months the answer choice is B. I took the 27 month data since this is policy year data and first retro adjustment for all the policies will be done only by 27 months.

9. AY 98 losses @Ultimate = $9000
CY 98 losses = 10000
No difference for years 1997 and prior
Difference = Adjusted losses - Actual Losses
= 9000 - 10000
= -1000
which is answer choice D.

What does everyone else think?

[Ugh]

For #7, I used the premiums at 27 months divided by the losses at 18 months. I'm pretty sure that's what the paper said to do.

For #9, I got the same answer.

[Avi]

Did anyone write down their numerical answers to any of the essay questions? I'll share the ones that I wrote down.

21.) 122,908
23.) 6,723,406
25.) 972,040
29.) a.) .3, .05, .05
b.) 1301
c.) redundant by 2,222,000
30.) 1,491,066
31.) 2002 - 16,461
2003 - 12,234
33.) a.) 2001, 3M
b.) 1/3
c.) 6.4M
34.) 5115
35.) a.) 754
b.) 770
c.) 755
36.) 17,965
39.) A-506
B-667
C-500
42.) a.) 243, 243
b.) 186, 300
43.) a.) .2024
b.) .4789
45.) d.) .125
47.) 1,875,000
49.) a.) 2100, .646
b.) 2168.4, .657

Behold! The anatomy of a failure!

20 - 129,477.10

21a - skip :duh:
21b - 1.2M
21c - 1.2M
21d - skip :duh: - but did write credibility weighting of a and 1.2M where Z = % rept or 1/2.2

22 - Development pattern and Loss severity trend

23a - 13% trend lat year vs. 6% overall and paid
23b - 6,455,436.40

24a - estimate all 1st adjustments
24b - estimate all 2nd and further adjustments
24c - changes to retro min/max & change to occurrence limit

25 - 972,225.60

26 - Same as last year :) Hurricane, event harder - waether is very complex to model but policy terms and conditions are clear. Pollution - Legal decisions and lack of data make it difficult to determine which policies and conditions apply, but actual loss simpler to moel.

27 - HAZARD - Risk of loss by peril such as fire
OPERATIONAL - Risk of loss by change in leadership
FINANCIAL - Risk of loss through poor (inadequate) pricing
STRATEGIC - Risk of loss through corporate embarrasment/reputation damage

28 - a) A loss reserve is an actaularially sound estimate of the costs to cover all incurred losses, both reported and unreported
b) due to uncertainty, a range can be valid
c) ALAE Reserve = a) but for incurred LAE
d) best reserve depends on financial requirements and underlying probabilities (yeah, that one is wrong :( )

29a - Chart - didn't copy down
29b - 1300.95
29c - 2,220,900
29d - I forgot what I wrote

30 - 1,860,945

31 - 28,689.79 (I had individual years, but forgot to write those down)

32 - fugeddaboudit

33a - b = 2001, c = 3,000,000
33b - 1/3 (ln(1.28)/ln(2))
33c - 6,419,952

34 - 5,115

35a - Didn't get a chance to write down
35b - Didn't get a chance to write down
35c - skipped :duh:

36 - 17,965.17

37 - :duh: GAAP, SAP, and two other's I'm too embarrased to mention :oops: :)

38a - The loss that may occur on this policy if a nat cat were to happen
38b - 1) Probable that loss will occur
2) Reasonable estimate of such losses
38c - No, b/c 1) loss is not probable, and b) there is no reasonable estimate in the future.

39 - A 600, B 600, C 375 <-- pretty sure incorrect

40 - :duh:

41 - see #40

42 - SAP 186/300
GAAP 243/243

43a - used ludwig methodology for average limit
43b - :duh:
43c - Bluffed somethingorother

44 - Have I used :duh: already? :oops:
Occurrence Primary Reinsuerer
1 140K 0
2 210K 100K
3 210K 450K
4 560K 700K

45a - Primary/R1/R2/R3
3.8M / 3.2M / 4M / 4M
45b - Primary 400K less Rein3 1.6M less
45c - Don't remember what I wrote
45d - 12.5%

46 - /sigh

47 - :crying: (1.125M?)

48 - .1411

49a - .739 - 25,449.70
49b - .556 - 23794.60

50 - Don't remember what I wrote.

[islas_del_maiz]

Let me try #44, becoz i'm getting different numbers. I did it by figuring out the reins share (this will be the reins's paid), then calc the primary's payment (the 70% factor applied to the WHOLE amt), and then deduct the reins share from taht amt to get the primary's actual paid.

I wrote down 140, 180, 75, 350 for occ 1, 2, 3, 4. Hope this sounds familiar to some!

By the way, for #30 the Johnson one, anyone got around $1,333,565.

[Jack Divine]

I got 1,333,564 for the Johnson one. I used the same approach as you describe for #44, but I only recorded the aggregate amounts on my exam (I really hope I listed it by claim on the actual test). I have 1.25M by the reinsurer, .745M by the primary, which matches the sum of your primary amounts.

[islas_del_maiz]

Thank you. I feel much better. For 44 though it did ask for amts to be paid for each claim by primary and reins.

Some more. For 45, I agree with the posted numbers. But for part b, I had cat reduced by 1.6 and primary by 3.2.

For 49, I had 2236.4 and 66.7% for part b.

Sadly I think I messed up on the exposure rating problems, again. I know that stuff!

I got 1,333,564 for the Johnson one. I used the same approach as you describe for #44, but I only recorded the aggregate amounts on my exam (I really hope I listed it by claim on the actual test). I have 1.25M by the reinsurer, .745M by the primary, which matches the sum of your primary amounts.

[Avi]

Hmmmm. I see what I did on #49. I calculated every year, and I think I may have written down the aggregate, although I calculated each year including 2002.

:swear: careless errors. Maybe I'll get partial credit :(

[KindGrind]

I hope you will, cause I did the exact same thing! :(

[Anita]

Avi,

On 33b., where does ln(1.28)/ln(2) come from? I just calculated the projected y's using 1/3 and 1/5 and compared with the actuals.

Thanks!

Side note: I thought the exam was more difficult than last year.

[Avi]

Assume b=2001 and c=3M

Now y = 2M*(x-2001)^p+3M

Use x=2003

y=sum(Inc Paid Losses) = 5,560,000

So we have 2*(2^p)+3 = 5.56

2^p=1.28

p = ln(1.28)/ln(2) ≈ .356 ≈ 1/3

[lee]

Assume b=2001 and c=3M

Now y = 2M*(x-2001)^p+3M

Use x=2003

y=sum(Inc Paid Losses) = 5,560,000

So we have 2*(2^p)+3 = 5.56

2^p=1.28

p = ln(1.28)/ln(2) ≈ .356 ≈ 1/3

For p=1/3 and p=1/5, I calculated each fitted values for all calendar years and the overall errors using the 1st error formula provided in the paper
sum(fitted value-actual value)^2/sum(fitted value)^2.

Then pick the one with less error, which is p=1/3.

I also think the exam is more difficult than last year.

[KindGrind]

I also think last year's exam was easier. I had 77% in 3 hours whereas in this fall's exam, I think I might get more around 62% in 3 hours 45 minutes.

[rafkooch]

Did anyone write down their numerical answers to any of the essay questions? I'll share the ones that I wrote down...

29.) a.) .3, .05, .05
b.) 1301
c.) redundant by 2,222,000
...

...
29a - Chart - didn't copy down
29b - 1300.95
29c - 2,220,900
...


For 29, part c.
While I also put down 2,222,000 for this question, it seems that it is wrong. Because the question asks whether the report year 2003 RESERVES are redundant or deficient. However, the first 70% of 22,000 claims (or 15,400) have already been settled at $660 per claim. So the actual remaining reserve by the insurance company is $1200*6600 claims, or $7.92 million. And the indicated remaining reserve should be the sumproduct of the remaining claims to be settled and their respective settlement values, which is $18.46 million. Therefore the reserves are deficient by $10.54 million. Does this seem right?

[Argh!]

No. The way you're doing it, you no longer have an average of $1,200 per claim because you're now saying average incurred is 660 for the first 70% and 1,200 for the remaining 30%. However, to look at it that way, you have to have something higher than $1,200 for the next 30%.

[rafkooch]

Okay, I got it.

For 35a, how did you get 754? I think that 754 is the modified expected loss, but the AY 2003 Ultimate should be 770?

Also, for 43 I'm not sure how you got your answers. If the question asked for pure premium, wouldn't that be a full dollar amount instead of a ratio?

Also, for 45b, when the reinsurance inures to the benefit of the reinsurance, then that portion comes right off the top. So in this situation, where the inuring reinsurance is 3M XS 12M, and the Cat Loss is $15M, then the inurer takes the first 3M and the remaining 12M is left for the remaining parties. So the Reinsurer C share would be 0.8*2M or 1.6M which is 2.4M less than before.

[Bobby Savoy]

For 35a, I got 775.

For 43, I also thought they were looking for dollar amounts rather than ratios, but I did put both in the answer (circled the dollar amounts, though).

For 45b, for reinsurer C, I got 2.4M less also (same logic as you have, rafkooch)