Actuarial Outpost
 
Go Back   Actuarial Outpost > Exams - Please Limit Discussion to Exam-Related Topics > SoA > Modules 6-8
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions


Reply
 
Thread Tools Search this Thread Display Modes
  #251  
Old 09-19-2019, 09:58 AM
TinyTim7's Avatar
TinyTim7 TinyTim7 is offline
Member
SOA
 
Join Date: Sep 2014
Location: Virginia
Studying for STAM
College: University of Illinois, Urbana-Champaign
Favorite beer: Root
Posts: 38
Default

Quote:
Originally Posted by ARodOmaha View Post
I used the modified Sharpe ratio in my first attempt, and CTE(90) in my second attempt, failing both. I hear what you guys are saying about using common sense and using the metrics as guides. But then how would you approach the sensitivity tests? It says to use the same risk/returns metrics. Do you have re-analyze every possible scenario in order to get the sensitivity? It would seem clearer with a formula approach like I was using.
For sensitivity testing, I used the asset class mix I recommended from task 2 and the given (default) assumptions as my base. The asset mix never changes, but the statistical measures will change when you vary an assumption. My sensitivity analysis was basically looking at the % change in statistical measures (the same measures I used in task 2) from the base assumption.
__________________
P FM MFE SRM FAP LTAM
STAM PA APC
Reply With Quote
  #252  
Old 09-19-2019, 10:30 AM
andrew.meier22 andrew.meier22 is offline
Member
CAS SOA
 
Join Date: Jan 2018
College: Rutgers Universit Alumni
Posts: 36
Default

It had been suggested to me to use the idea of an efficient frontier to pick an asset class mix. I understand how I would use that to choose the mix but I’m not seeing how you could really do a sensitivity analysis using this? I guess you would just have to plot the new return and SD on the same play as the initial scenarios and see Where it lies but there’s really no quantitative measure of that change.
Reply With Quote
  #253  
Old 09-22-2019, 11:33 AM
mnm4156 mnm4156 is offline
Member
SOA
 
Join Date: Oct 2015
Posts: 38
Default

Can someone please explain what using the CTE 95 metric for determining the cost per employee would tell us in the context of the CDEF situation? Let's say that using the CTE 95 metric would result in a cost per employee of 800. So a tax amount of 800 would sufficiently cover future liabilities with 95% certainty? or is that what Var 95 tells us?
Reply With Quote
  #254  
Old 09-22-2019, 01:01 PM
ARodOmaha ARodOmaha is offline
Member
SOA
 
Join Date: May 2016
Location: Omaha, NE
College: University of Nebraska (alma mater)
Favorite beer: Captain Morgan
Posts: 204
Default

Quote:
Originally Posted by mnm4156 View Post
Can someone please explain what using the CTE 95 metric for determining the cost per employee would tell us in the context of the CDEF situation? Let's say that using the CTE 95 metric would result in a cost per employee of 800. So a tax amount of 800 would sufficiently cover future liabilities with 95% certainty? or is that what Var 95 tells us?
CTE(95) is the mean cost per employee in the top 5% of scenarios. Since we don't know the exact distribution, you could only say it would be sufficient in OVER 95% of scenarios. Your statement would apply more to VaR(95).
__________________
P FM MLC MFE C PA FAP APC
Reply With Quote
  #255  
Old 09-25-2019, 06:03 PM
ReturnStudent ReturnStudent is offline
Member
SOA
 
Join Date: Dec 2015
Location: Chicago
Studying for PA
College: UNL Alumni
Posts: 210
Default

Quote:
Originally Posted by Sir Issac View Post
Original submission I only used mean, cte and var as risk/return metrics and said minimize all but like I said in my post they are not really a risk/return metric. And I suggested minimizing the mean/cte in my second

Basically if two portfolios have a similar cte the one with the lower mean has better risk/return tradeoff
Would someone please explain to me why this is the case?
I'd think the other way around was correct because:
Mean / CTE = expected return rate over risk.
Given 2 portfolios with the same risk value (CTE in the denominator), the one with higher mean (expected rate of return) would have better risk/return trade off.

Thanks!
Reply With Quote
  #256  
Old 09-25-2019, 06:23 PM
DjPim's Avatar
DjPim DjPim is offline
Member
SOA
 
Join Date: Nov 2015
Location: SoCal
Posts: 658
Default

Quote:
Originally Posted by ReturnStudent View Post
Would someone please explain to me why this is the case?
I'd think the other way around was correct because:
Mean / CTE = expected return rate over risk.
Given 2 portfolios with the same risk value (CTE in the denominator), the one with higher mean (expected rate of return) would have better risk/return trade off.

Thanks!
To answer your question, if two portfolios have the same CTE, then they are similar in risk, correct. As in, the bad-case scenarios are similar. Remember that in this case, we want lower numbers, because this is a cost imposed. So, the lower the mean required contribution, that means you got a higher return, and therefore it is better.

Worth repeating though that pretty much all feedback received is to not use a ratio.
__________________
Quote:
Originally Posted by Dr T Non-Fan View Post
"Cali" SMH.
Reply With Quote
  #257  
Old 09-25-2019, 07:29 PM
mnm4156 mnm4156 is offline
Member
SOA
 
Join Date: Oct 2015
Posts: 38
Default

how do we know that the mean measures return only and CTE measures risk only? i understand CTE is a risk measure, but its not clear to me why the mean would be a return measure.
Reply With Quote
  #258  
Old 09-25-2019, 07:40 PM
DjPim's Avatar
DjPim DjPim is offline
Member
SOA
 
Join Date: Nov 2015
Location: SoCal
Posts: 658
Default

Quote:
Originally Posted by mnm4156 View Post
how do we know that the mean measures return only and CTE measures risk only? i understand CTE is a risk measure, but its not clear to me why the mean would be a return measure.
The higher the return on your investments, the less money you need to collect as a tax. Mean will give you a rough idea of an expected required tax, and if that mean is low, that means your returns must have been high.
__________________
Quote:
Originally Posted by Dr T Non-Fan View Post
"Cali" SMH.
Reply With Quote
  #259  
Old 09-25-2019, 09:11 PM
mnm4156 mnm4156 is offline
Member
SOA
 
Join Date: Oct 2015
Posts: 38
Default

Quote:
Originally Posted by DjPim View Post
The higher the return on your investments, the less money you need to collect as a tax. Mean will give you a rough idea of an expected required tax, and if that mean is low, that means your returns must have been high.
i understand that but can't you also look at it both ways? its measuring how risky the asset portfolio is because its calculating the mean tax contribution that is needed to offset the level of risk of the portfolio. i feel like thats why the Q refers to the metrics as risk/return metrics and not risk and return metrics
Reply With Quote
  #260  
Old 09-26-2019, 02:52 PM
DjPim's Avatar
DjPim DjPim is offline
Member
SOA
 
Join Date: Nov 2015
Location: SoCal
Posts: 658
Default

Quote:
Originally Posted by mnm4156 View Post
its measuring how risky the asset portfolio is because its calculating the mean tax contribution that is needed to offset the level of risk of the portfolio.
I think you're confused about what a 'risk' is.

Theoretically, let's say your asset portfolio is 1/3 each in treasury, bonds, and equities. You simulate annual returns from each asset, and somehow each of your 1 million simulations tells you that you get the exact same return every time. If I were to ask you what you expect your return to be on the next simulation, I imagine you'd be pretty sure it's the same as the last million, as in, low risk of having adverse outcome. In this question, it doesn't matter how much your returns are, because I'm asking about the risk of it being adverse. In other words, there is 0 volatility.

Now let's say none of your 1 million simulations are the same. Some say you lose all your money, some say you make 1000% return some years, etc etc--it's all over the place. Averaging these simulations up, you find your "average" return. Now if I ask you what you think the exact return will be on your next simulation, you'd probably say there's no way to know, it varies too much. You could tell me the average, which gives me a ballpark of where your distribution might be centered, but you wouldn't be confident the next simulation would be close to the average. Therefore, high risk, high volatility.

ETA: The average you do find in scenario 2 still tells you something about your return, just not the level of risk associated with it. Easy to picture a bell curve, and this average tells you where your distribution is centered, but doesn't tell you the weight of the tails.
__________________
Quote:
Originally Posted by Dr T Non-Fan View Post
"Cali" SMH.
Reply With Quote
Reply

Tags
cdef, fa task 2, final assesment

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


All times are GMT -4. The time now is 06:30 AM.


Powered by vBulletin®
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
*PLEASE NOTE: Posts are not checked for accuracy, and do not
represent the views of the Actuarial Outpost or its sponsors.
Page generated in 0.31290 seconds with 11 queries