- This topic has 3 replies, 2 voices, and was last updated 8 years ago by
.
- Posts
I’m just finishing off revising this topic, but, there’s one aspect that’s a little puzzling. Often the question will tell you the moving average range to work from- i.e 5 or 3 for example (which is fine and presents no problem) but if you’re trying to work out the trend and seasonal variations for four quarters over a year, won’t a 3 range moving average be better than a 4, as you’re trying to compare the 4 seasonal quarters?
Additionally, if you’re not told in the question what the moving average range is, say for ten years with one entry for say ” total sales volume” each year would you just begin your average calculation for the start of the third year down, but again what moving average range do you select if you’re not told, 3, 4..?
.
Firstly, an exam question will always make it clear over which periods the seasonality occurs.
Secondly, in practice, you would have to find out first over which periods the seasonality occurred – probably by plotting on a graph and seeing if there was a pattern, and if so over which periods. It is not a question of one range being better that another. (And if the seasonality does occur quarterly (as is usually the case in exam questions) then to get the average you would divide by 4 because there are four quarters – there would be no logic in dividing by 3.
Finally, seasonality is only ever short-term fluctuations – generally over a year. You would never consider yearly fluctuations over a ten year period – that would not be seasonality.
Thanks, Yes 4 is logical and makes sense, but, some of the Kaplan questions have asked for a moving average over 3 or 5, which isn’t perhaps what’s most used in the real world.
Yes, I get your point about the seasonality pertaining to a short time range (a few years at most)….I was thinking if cyclical trends which you said in another post can’t really be modeled very accurately.
As I wrote before, exam questions will always make it clear what periods to use (I agree that 3 and 5 are not very realistic, but obviously in the middle of an exam you do what is asked for 🙂 ).
For cyclical trends you would not use the same approach because they cannot really be modelled (and obviously you could not be asked to deal with them in the exam).
- You must be logged in to reply to this topic.