Dependent variable is function of itself

#1
I have a model: qty=a.price+....(other IVs)
Price is calculated as sales_amount/qty.
Hence for a customer, if we aggregate data of a certain time-period (say month), we get:
q1+q2+q3+...=a.(s1+s2+s2+...)/(q1+q2+q3+...)+....(other IVs)
where s(i) denotes sales_amount.
Will there be any problem because of this?
 

Mean Joe

TS Contributor
#2
if we aggregate data of a certain time-period (say month)
You do not need to aggregate data.

Does the price really change that much day-by-day? If so, I probably would not aggregate data.

The part where you calculate (s1+s2+s3+...)/(q1+q2+q3+...) really seems like something you should not do. Looks like I would instead calculate s1/q1, s2/q2, s3/q3, ... ie, I would not be aggregating data.
 
#3
Ok so let me simplify the problem. Instead of aggregating, if we just keep:
q1=a.s1/q1+....(other IVs)
Will this create a problem (since the DV is a function of itself)?