Actuarial Outpost > SoA CaN-Do: Task 7
 Register Blogs Wiki FAQ Calendar Search Today's Posts Mark Forums Read
 FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions

 Upload your resume securely at https://www.dwsimpson.com to be contacted when new jobs meet your skills and objectives.

#11
08-08-2012, 02:00 PM
 ebeebs Member SOA Join Date: Jul 2008 Location: New England Posts: 2,989

Quote:
 Originally Posted by KevinR Btw can I just ask one thing, are the actual labor costs higher or lower than the expected labor costs (based on the actual gold/ton yield given)? My expecteds were lower than the actual, but a friend of mine said his expected labor costs were higher than the actuals. I think this particular result should be similar for everyone, as it doesn't depend on the random values at all.
Labor is linked to price of gold and number of shifts. Gold prices are simulated and I don't remember what info the macro randomly generated but if the gold price simulation is one of them then expected labor will be different.

Also, my actual was greater than expected.
#12
08-08-2012, 02:04 PM
 ebeebs Member SOA Join Date: Jul 2008 Location: New England Posts: 2,989

Looking at my paper, all I changed was the inflationEXCHANGE RATE** and labor cost a for 1 and 2 shifts.

Last edited by ebeebs; 08-09-2012 at 06:46 PM..
#13
08-08-2012, 03:33 PM
 DonchaKnow Member SOA Join Date: Feb 2008 Favorite beer: Stella Posts: 190

Quote:
 Originally Posted by ebeebs Looking at my paper, all I changed was the inflation and labor cost a for 1 and 2 shifts.
yea those are sufficient, but more can be inferred from previous analysis
#14
08-08-2012, 03:41 PM
 ebeebs Member SOA Join Date: Jul 2008 Location: New England Posts: 2,989

Quote:
 Originally Posted by DonchaKnow yea those are sufficient, but more can be inferred from previous analysis
such as?
#15
08-08-2012, 03:50 PM
 insomnie Member Join Date: May 2006 Posts: 80

The actual decay rate is a single number.

Before opening the mine, the rate is unknown. What is assumed and what we are told not to question, is that the actual rate is normally distributed. This distribution is used to model the rate.

After opening the mine, we find out the actual rate. This known rate should be used for future projections.
#16
08-08-2012, 04:04 PM
 ebeebs Member SOA Join Date: Jul 2008 Location: New England Posts: 2,989

Quote:
 Originally Posted by DonchaKnow yea those are sufficient, but more can be inferred from previous analysis
Quote:
 Originally Posted by ebeebs such as?
The only other things you could really argue changing are non-labor costs and inflation. I argued against changing them in my submission.
#17
08-08-2012, 05:18 PM
 IIRC Member Join Date: Apr 2009 Posts: 397

Quote:
 Originally Posted by insomnie The actual decay rate is a single number. Before opening the mine, the rate is unknown. What is assumed and what we are told not to question, is that the actual rate is normally distributed. This distribution is used to model the rate. After opening the mine, we find out the actual rate. This known rate should be used for future projections.
Yeah but, in light of the Expected being N(.945, .02), is it all that surprising for the Actual to be what it is?

And it does not seem to be part of the Scenarios and Result tab, as mentioned.

IMO, granted I havn't passed yet.
#18
08-08-2012, 06:23 PM
 insomnie Member Join Date: May 2006 Posts: 80

Quote:
 Originally Posted by IIRC Yeah but, in light of the Expected being N(.945, .02), is it all that surprising for the Actual to be what it is? And it does not seem to be part of the Scenarios and Result tab, as mentioned. IMO, granted I havn't passed yet.
It's not a surprising result... however, the decay rate is used to project ending cash value. Why not use the actual rate, rather than the original estimate, to improve the quality of the projections?
#19
08-08-2012, 08:49 PM
 lucas_m Member SOA Join Date: Nov 2006 Studying for FSA modules Favorite beer: Alexander Keiths Posts: 256

Quote:
 Originally Posted by insomnie It's not a surprising result... however, the decay rate is used to project ending cash value. Why not use the actual rate, rather than the original estimate, to improve the quality of the projections?
The decay from assays is a stochastic process.
The question asks what assumptions should be changed not to change or improve the model. Remember, the purpose of the model is to simulate gold prices and decays to project mining cashflows. If you use a constant decay rate then the model is useless. For example, if you simulate stock return using a lognormal distribution and you observe a return of 5%. Are you going to say, I will use 5% for future projection because this is actual experience?
The correct approach would be to calibrate the new mean and standard deviation of the normal dist using the actual decay data. But again, I don't think this is required. Just change the assumptions you juge should be changed within the assumption tab in the model spreadsheet and don't overthink this.
#20
08-08-2012, 10:43 PM
 KevinR Join Date: Jun 2011 Posts: 16

Quote:
 Originally Posted by ebeebs Looking at my paper, all I changed was the inflation and labor cost a for 1 and 2 shifts.
How did you determine the new inflation rate? I tried matching the other costs cash flows by quarter (using the actual gold/ton and number of shifts, i.e my gold/ton and number of shift columns aren't driven by formulas anymore but rather I simply pasted the given values into them).

I couldn't get them to match at all no matter what inflation I used. I mean, even the first quarter other costs already had a big difference and the first quarter has got nothing to do with inflation.

so in the end all I changed were the exchange rate and the labor cost formula for two shifts. Due to the very small data sample, I couldn't determine the the labor cost formula for 1 shift, and so my stop gap measure for the one-shift labor cost was simply to multiply the two-shift cost by 0.5 (this gives a pretty close approximation of the actual labor costs).

 Tags can do, can-do