how we calculate the --->adjustments<
in time series?
in time series?
ACCA Forums
PMTime Series
The point of adjustments is to make total variation of final estimates 0.
After calculating quarterly average variations (of all years) you sum up to TOTAL and this is what needs to be adjusted in total. So if TOTAL variation is 100 the adjustment will be in TOTAL (-100) and for period divided by number of periods. So if they were months you would divide by 12 to give equal amount of adjustment to each period.
Q1 + Q2 + Q3 +Q4 = TOTAL 100
Adjustment for each period 100/4 = 25
This was additive model.
With multiplicative model the average variation TOTAL (first line) will be just around the number of periods. So if u have quarters it will be around 4.0 for instance 4.08. If months it will be around 12 for example 11.93
So the adjustment will be (4.00 - 4.08)/4 = -0.02 for each quarter
or (12 - 11.93)/12 = very small :) for each month.
so after adjustments the final variation TOTAL will be 4 (for quarters) / 12 (for months).
awww i "wrote" a table but it mixed up after posting so had to resign from it.
I hope this helps anyway.
After calculating quarterly average variations (of all years) you sum up to TOTAL and this is what needs to be adjusted in total. So if TOTAL variation is 100 the adjustment will be in TOTAL (-100) and for period divided by number of periods. So if they were months you would divide by 12 to give equal amount of adjustment to each period.
Q1 + Q2 + Q3 +Q4 = TOTAL 100
Adjustment for each period 100/4 = 25
This was additive model.
With multiplicative model the average variation TOTAL (first line) will be just around the number of periods. So if u have quarters it will be around 4.0 for instance 4.08. If months it will be around 12 for example 11.93
So the adjustment will be (4.00 - 4.08)/4 = -0.02 for each quarter
or (12 - 11.93)/12 = very small :) for each month.
so after adjustments the final variation TOTAL will be 4 (for quarters) / 12 (for months).
awww i "wrote" a table but it mixed up after posting so had to resign from it.
I hope this helps anyway.
Thank u so much!! it is really helpful for me... I have one more question (maybe a stupid one) . How I will know if i haveto use additive model or multiplicative model in exam?
In my personal opinion the Examiner should say which method you are supposed to use.
Well generally the trend u will find is either increasing or decreasing (not both).
In this case multiplicative model is better than additive because seasonal variations are also increasing or decreasing with trend.
The additive model is simply adding same value of adjustment no matter if the variation is positive or negative. So as I understand the problem with additive model is that if suppose the calculated value of adjustment is negative, and the seasonal variation is negative - by making this negative adjustment you are making this varation even more negative (even more far from trend).
Multiplicative model has proportional seasonal variations and this overcomes the above problem (you dont have -/+). And the "proportional" adjustment is made.
However additive is much more simple to use.
I'm not 100% sure about my approach (understanding) so maybe someone should correct me. I'm not trying to go more deep into this because
I found only one question worth 5 marks on this topic of all questions in revision kit, and it was indeed explaining why one is better than another. So not a big deal i guess...
Well generally the trend u will find is either increasing or decreasing (not both).
In this case multiplicative model is better than additive because seasonal variations are also increasing or decreasing with trend.
The additive model is simply adding same value of adjustment no matter if the variation is positive or negative. So as I understand the problem with additive model is that if suppose the calculated value of adjustment is negative, and the seasonal variation is negative - by making this negative adjustment you are making this varation even more negative (even more far from trend).
Multiplicative model has proportional seasonal variations and this overcomes the above problem (you dont have -/+). And the "proportional" adjustment is made.
However additive is much more simple to use.
I'm not 100% sure about my approach (understanding) so maybe someone should correct me. I'm not trying to go more deep into this because
I found only one question worth 5 marks on this topic of all questions in revision kit, and it was indeed explaining why one is better than another. So not a big deal i guess...
Sign in to reply to this topic.