RE: st: Calculating cumulative value over several observations

[Nick Cox <n.j.cox@durham.ac.uk>]

Note that my earlier 

sum(((Seller == `j') * Price) - ((Buyer == `j') * Price))

can naturally be reduced to 

sum(Price * ((Seller == `j') - (Buyer == `j')))

and a similar simplification can be applied to Clyde's independent solution. 

Nick 
n.j.cox@durham.ac.uk 

Clyde Schechter
===============

I think this will get you what you want:

levelsof buyer, local(buyers)
levelsof seller, local(sellers)
local participants: list buyers | sellers
sort market period observation

foreach p of local participants {
	by market period (observation): gen M`p' = sum(price*(seller==`p')) -
sum(price*(buyer==`p'))
	by market period (observation): gen A`p' = sum(buyer == `p') - sum(seller
== `p')
}

Note:  If your buyers and sellers constitute the integers between 1 and P
for some P, then you don't need to do the -levelsof- commands and you can
replace the foreach loop with a forvalues p = 1/P loop.

By the way, just out of curiosity, if everybody starts out with no money
and no apples, how do they make these trades?

Nick Cox
========

forval j = 1/6 {
	bysort Market Period (Obs): gen M`j' = sum(((Seller == `j') * Price) - ((Buyer == `j') * Price)) 
	by Market Period (Obs): gen A`j' = sum(((Buyer == `j') - (Seller == `j')) } 

Palan, Stefan 

In my dataset every observation is one trade between a buyer and a seller. Each person is uniquely identified by an ID and can buy and sell (so the same ID will be a buyer in one observation and a seller in another observation). In fact, some example observations would look like this:

Observation Market Period Price Seller Buyer
1           1      1      100   4      2
2           1      1      105   4      6
3           1      1      103   2      5
4           1      1      107   6      4
5           1      2      89    3      1
6           1      2      90    4      3
7           2      1      79    3      2

Each transaction here is for, say, 1 apple. Now I would like to calculate, observation by observation, each person's holdings of money and apples, resetting with every new market and period. The result could look as follows (assuming every individual starts out with no money and no apples, M# stands for money of person #, A# stands for apples of person #):

Obs Market Period M1    M2    M3    M4    M5    M6    A1  A2  A3  A4  A5  A6
1   1      1      0     -100  0     100   0     0     0   1   0   -1  0   0
2   1      1      0     -100  0     205   0     -105  0   1   0   -2  0   1
3   1      1      0     3     0     205   -103  -105  0   0   0   -2  1   1
4   1      1      0     3     0     98    -103  2     0   0   0   -1  1   0
5   1      2      -89   0     89    0     0     0     1   0   -1  0   0   0
6   1      2      -89   0     -1    90    0     0     1   0   0   -1  0   0
7   2      1      0     -79   79    0     0     0     0   1   -1  0   0   0

Unfortunately, I don't know how to efficiently add up the information in this way. Can somebody please help me with reshaping my data to get this output?


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/