how can I generate an auction dataset that has some variation in the number of bidders?

[Li Zhang]

Hi All,

I would like to generate an auction dataset that has some variation in the number of bidders across auctions. For example, in the following example, we have 9 auctions, and maybe in auction 1 we have 7 bidders, auction 2 we have 5 bidders. Also, in each auction, the bids are drawn from a uniform distribution.
I know how to generate a balanced one(each auction has the same bidders) but an unbalanced one. Any help would be appreciated ! Thanks

auction_id	bidder1	bidder2	bidder3	bidder4	bidder5	bidder6	bidder7	bidder8	bidder9
1
2
3
4
5
6
7
8
9

[Mike Lacy]

If want to use the wide format (perhaps not the best choice), you can leave some bidder values missing.

clear
set obs 10
local maxbidder = 15
gen int auction_id = _n
gen int nbidder = ceil(`maxbidder' * runiform()) // one possibility
forval i = 1/`maxbidder' {
   // for example
   gen bidder`i' = rnormal(10,2) if (`i' <= nbidder)
}

Or, you might do better with the long format:

clear
set obs 10
gen int auction_id = _n
local maxbidder = 15
gen int nbidder = ceil(`maxbidder' * runiform()) // one possibility
expand nbidder
bysort auction_id: gen int bidder = _n
// for example
gen bidder = rnormal(10,2)

[Li Zhang]

Hi Mike,

Thank you so much. That exactly what I am looking for. I really appreciate it.

All best,
Li