April 2007

[scarlets]

Hi there
Did anyone else sit this one?
I thought it was OK, could have been much harder, however as ever they make sure there's 2-3 big questions at the end that you have to rush through like a nutcase. 14 questions on thsi paper.

Personally I didn't work out values for the Q14 (18 mark)- only wrote down the method etc. That's so frustrating. Let's hope they take the knowledge presented on the first 13 questions warrants a pass. I hope they value knowledge presented in reasonable time before failing people for not being able to finish.

 

[Monkey_Mike]

Hiya,

Yeah I sat this, I was generally just about OK for time. Managed the last Q, but had skipped an earlier one, (the pension fund one), so went back to that at the end and rushed it a bit.

I thought the Covariance question didn't make sense, (rushed this at the end too), It seemed to have a typo somewhere, and I think it should have been Cov(X,Y) and not, Cov( E[X], E[Y] ) like it showed..

A few people I'd spoken too afterwards had used dependent rates for the last Q (or second to last), when I think it should have been independent rates...

It started off OK ish...

 

[scarlets]

Hiya,

Yeah I sat this, I was generally just about OK for time. Managed the last Q, but had skipped an earlier one, (the pension fund one), so went back to that at the end and rushed it a bit.

I thought the Covariance question didn't make sense, (rushed this at the end too), It seemed to have a typo somewhere, and I think it should have been Cov(X,Y) and not, Cov( E[X], E[Y] ) like it showed..

A few people I'd spoken too afterwards had used dependent rates for the last Q (or second to last), when I think it should have been independent rates...

It started off OK ish...

yes the Cov thing looked wrong to me also for the same reasons, this really put me off actually, hope they realise how things like this can waste people's time. That's 10 minutes more I could have spent on Q14...

 

[AlphaCharlie]

Yes, there were 2 errors in Question 10 of the CT5 paper today.

Firstly, and fairly innocently, there was a typo, which missed the "colon n" subscript of a term assurance, so that it was just an A[x] with a 1 above the 'x'.

While this was forgivable, there was an appalling mistake where they asked for the covariance of the difference of two assurances, written in "A" form, rather than the covariance of the underlying random variables (which was required to prove the result in the question).

Given the numerous patronising comments over the years in the Examiners' Reports about candidates failing to appreciate the difference between a random variable and its expected value, this is particularly irritating. Thankfully, I stated my inference as to what they meant to ask and proceeded, but it's quite possible that other candidates either got bogged down in trying to evaluate something incorrect, or got put off and didn't attempt it at all.

In terms of the rest of the paper, I think it was quite fair. Some very unusual bits, e.g. the annuity payable from birth (presumably), which provided a death benefit as well as a benefit paid out every 5 years (did I understand that correctly?)

I've just seen the posting about dependent decrements. If this was what was required, I've got it a bit wrong: I used the rates independently.

In fact, I've just realised that, in working out the profit margin, I completely ignored surrender...oops! Also got some funny (presumably wrong) numbers, indicating a negative profit margin. Going to hope for lots of method marks on that one! :-D

 

[Monkey_Mike]

I'm hoping my marker, and the examiners, are called Mr M Ethod (M stands for Mark) and Mrs. Method !

It was OK, a lot of fiddly bits, but hey, this is CT5

The rates should have been independent, since surrenders were only possible at the end of the year, they wont affect the death probs

 

[Monkey_Mike]

and yeah, I did that one (Q5) from 0 to infinity seeing as they didnt specifiy an age, I just did it from birth

The one with 2 bonus payments took me a while to figure out, got the simple one kind of right, (except I had problems with (IA)[x]+1 (seeing as its not in the tables.!! )

Also The compound one I got it working out to be 50,000/v^0.5 *1.04

for the SA and bonuses.. because the 1.04 and the v's cancelled out, so this just left sum( k=0 to inf ) kp[x]*q[x]+k which I got as 1?!? (probably wrongly)

Alpha Charlie, you sound like you've done pretty well overall

Was pleased at the joint life one with a guaranteed term, spent last night doing one exactly the same (pretty much)

 

[AlphaCharlie]

and yeah, I did that one (Q5) from 0 to infinity seeing as they didnt specifiy an age, I just did it from birth

The one with 2 bonus payments took me a while to figure out, got the simple one kind of right, (except I had problems with (IA)[x]+1 (seeing as its not in the tables.!! )

Also The compound one I got it working out to be 50,000/v^0.5 *1.04

for the SA and bonuses.. because the 1.04 and the v's cancelled out, so this just left sum( k=0 to inf ) kp[x]*q[x]+k which I got as 1?!? (probably wrongly)

Alpha Charlie, you sound like you've done pretty well overall

Was pleased at the joint life one with a guaranteed term, spent last night doing one exactly the same (pretty much)

Hmmm...I didn't get any IA[x]+1...I just had IA[x]. I've probably messed up, though. Again, I'm hoping for method marks!

With regard to the compound bonus, yes, I got that it comes out to 1! This is completely intuitive (though I, like you, confirmed by formula): if there is no interest, the value of a whole-life assurance is precisely 1, since the death payment of 1 is guaranteed to happen at some point and there is no discounting. I got the expression to be 50,000v^0.5, I think. Maybe I've got that the wrong way round, though, and it should be 50,000/v^5.

You, too, sound like you've done well. Let's hope we both get it!

Incidentally, here is the latest from ActEd tutor, Julie Lewis, on the mistakes in Q10:

"I have contacted the staff actuary about these problems. She has told me
that she has spoken to the examiners and it will be taken into account
during the marking process."

Must get on with other stuff now...exam tomorrow --- eek!

Good luck!

 

[Monkey_Mike]

I think I got (IA)[x]+1 just because the bonus didn't vest until the end of the year, maybe I should have used 48,000A[x] + 2,000(IA)[x] (1) (A[x] with a bar on)

but spur of the moment I changed it to 50,000A[x] + p[x] 2,000(IA)[x]+1

I'd have to see the question again to figure out my skewed logic, something to do with bonus vesting at the end of the year so didn't think to use (1) because (IA)x would be paid at the end of the year and if they died in the first year they would only get 48,000 using (1) .. I'm confusing myself now just thinking about it

Good luck tomorrow

 

[Michael_14]

hi everyone,

i sat this paper yesterday too. I noticed the error in question 10 with regards to the covariance. I also wondered if the A_x^1 notation was an error and fudged my answer to that a bit because of it. It was a bit disappointing to have the error wasn't it? After all the hard work that the students have put in it seems a little disrespectful to us to make a mistake like that, but oh well that's life i guess.

I finished the thing, but similar to AlphaCharlie, I forgot the maturity payment in the final question - whoops. I used dependent decrememnts for death since the surrenders happened at EOY, but the dependent surrender decrememnts i used we a little lower than 5% due to deaths. All in all the last question was quite a long one and im sure many people have less than perfect answers due to time pressures.

I thought the paper was reasonable, though i agree it was pushed for time compared to past papers. I was glad the pensions question wasn't a super long one. I think its silly when they test the pensions part with complicated questions, it doesnt serve additional purpose in demonstrating understanding of the concepts.

I got thrown on the very first part of question 1 which was annoying! I just drew a blank on it! nerves i guess - and at the end i didnt have time to go back to it.

Otherwise i just calculated EPV after EPV after EPV! There were lots of them in that paper!

well done to everyone...
Michael

 

[Benny]

When there's an error in the question paper what happens with the marking of it?

I've heard it may even be ignored altogether? Is this realistic? That would be superb from my point of view!

 

[Monkey_Mike]

They'd mark our answer wrong if we did that, the marks should be scrubbed off the paper for a mistake this glaring

it wasn't just like a "could be interpreted this way or that" it was just wrong

 

[omniscientbeing]

I have to agree that the error was damn irritating, especially in an exam when you're very time conscious.

Does anyone remember how many marks that question with the gauranteed annuity for 5 years, and then the reversionary part was worth? The question that was repeated from a few years back. Out of curiosity i checked, and I think it was just back in september 2005, and then it was worth 9 marks.

I seem to remember it only being worth around 6 marks on our paper, because I was feeling pretty happy when I was doing it thinking that I had some easy marks, and then I was disappointed when I realised it had still taken me too long to do considering the small amounts of marks i was getting.

 

[Monkey_Mike]

Yeah I also did that one the night before the exam on the old paper,

it was definately worth less this time around, with just as much work

 

[scarlets]

Yes it's inexcusible to have an error like that on the paper.

 

[Louisa]

Error in Q10

I was also very unimpressed by the error in the paper, though lucky enough not to be thrown by it.

I don't believe there is any way this can fairly be "taken into account" in marking the paper. There's no way the examiners can tell how much time a student has wasted in trying to understand this question.
However I'm sure they'll do their best.

Does anyone know if there is a history of issuing official apologies for such errors? I think that is important here. Apology seemed notably lacking from the response reported by AlphaCharlie. (But maybe it got lost in the chinese whispers!)

I'm also wondering if we need some kind of possibility, as we had in A-levels and uni exams, of being able to ask the actuarial invigilator about any ambiguity or error in the exam paper. I seem to recall at uni this was permitted in the first half-hour of the paper.
Any thoughts?

 

[Little Miss Actuary]

I'm also wondering if we need some kind of possibility, as we had in A-levels and uni exams, of being able to ask the actuarial invigilator about any ambiguity or error in the exam paper. I seem to recall at uni this was permitted in the first half-hour of the paper.
Any thoughts?

Yes, at uni we could point out anything we thought was a mistake (at any time during the exam) and it was checked out by the chief examiner. If they agreed then a correction was written on a large black board at the front in each exam room. I don't see why a similar logic couldn't be applied here, although it would involve someone phoning each exam centre.

 

[scarlets]

I don't believe there is any way this can fairly be "taken into account" in marking the paper. There's no way the examiners can tell how much time a student has wasted in trying to understand this question.
However I'm sure they'll do their best.

Exactly. Some people might have spotted it was wrong as soon as they read the question. Others (me included) would have started to do the question and use it without considering the possibility it was an error. Others would have tried to "make it work". And so on. Given how precious time is in these exams, 10 minutes wasted here is 10 minutes not picking up marks on other questions and we could be talking about a crucial 5-10 marks here.

 

[TheChap]

Why the time pressure anyway?

I really do hate to be a moaner but the error aside(during the exam I sat there trying to convince myself I had done it right and that the examiner had clearly made an error) I am yet to get a good reason that exams such as CT5 are as time pressured as they are.

I am annoyed that I was unable to finish this exam, seeing as (again, error aside) I barely spent any time thinking, I didn't check over my answers; in essence I did not spend any unproductive time.

The only crime I may have been guilty of is being thorough with the way I write mathematically (I define many of my parameters explicitly). This is the way I was taught to do mathematics and I do not believe it makes me a bad mathematician as it is not an area of study that usually requires unnecessarily speedy delivery.

My understanding of CT5 when I sat down for the exam on Tuesday was certainly at the level necessary to pass an exam on the material. However, I may very well have failed because I tried to ensure my work was well laid-out (which might make me a bad actuary, I suppose).

If the purpose of the exams, as most qualifieds will stress time and time again, is to produce the most able actuaries, I am at a loss as to why we are not given enough time to finish all the work comfortably, and maybe even (perish the thought!) check over our answers? Surely, this means that many able students will fail even thought they have met any job requirement necessary (being able to find the covariance of a whole of life and a term assurance in six minutes flat is, in my opinion, not a key skill necessary for the profession... and I have my doubts the vast majority of qualifieds would be capable of doing so).

Perhaps I am totally misguided?

 

[examstudent]

is the covariance:

term assurance (rate i^2 + 2 i) - (term assurance at rate i ) * (whole of life assurance at rate i )

was the question something like find variance (term assurance - whole of life assurance)
and then just use above result for covariance

 

[scarlets]

TheChap, great post. I feel the same way. Fine if they want to make the exams hard, but I'd rather they made the questions harder/more involved than give us too much to do requiring us to rush and not have adequate time to think things through. I don't think I check over anything I do there simply isn't time.

Like in the CT6 exam I worked through one 9 mark question and didn't get an answer that looked very sensible. So this puts me in a difficult dilemma- do I have another go at the question? Do I check my workings to spot a silly error? Because if I spend any more time on that question that is time I can't spend picking up marks elsewhere and if I can't finish the paper I may live to regret that decision.