Bidding in FirstPrice Sealed Bid Auctions Without Feedback Information: The Interaction Effect betwe

1
Bidding in First-Price Sealed Bid Auctions
Without Feedback Information The Interaction
Effect between Bidding and Market Size

• Tibor Neugebauer
• Universität Hannover

Prepared for the ESA World-Conference, 2007 LUISS
Rome
2
Paper outline

• Introduction
• Experimental study B human competitors/ Blind
  interaction
• Experimental study C Computerized competitors/
  feedback on winning
• Bidding function of market size N
• Further results
• Concluding remarks

3
1 Auction at Christies
4
1 Theory of first-price sealed-bid auction in the
independent private value model

• Research question Are risk neutral Nash
  equilibrium bids and expected prices good proxies
  for bids and prices if N becomes larger?

• RNNE risk neutral Nash equilibrium bid (Vickrey
  1961 JF)

• Expected high bid (x1 high value)

5
1 Related literature

• Evidence on overbidding relative to the risk
  neutral Nash equilibrium in auctions with
  feedback information on high bid
• Kagel 1995 HEE
• .
• Evidence on underbidding relative to the risk
  neutral Nash equilibrium in auctions without
  feedback information on high bid
• Neugebauer and Selten 2006 GEB
• Neugebauer and Perote 2007 ExEc

6
2 Experimental study B Design

• Interactive BLIND experiment without information
  feedback
• n human competitors
• xit U0, .., 100 iid, i 1, .., n, subjects
  value
• n 3, 5, 7, 9, 14, treatments
• software toolbox zTree
• Observations 110 first-year student subjects x
  50 bids
• No information about payoff or competitors bids
  after each auction. Only at the end of the
  experiment, payoff is revealed.

7
2 Difference of bid-value ratio and RNNE ratioN
3
8
2 Difference of bid-value ratio and RNNE ratio N
5
9
2 Difference of bid-value ratio and RNNE ratio N
7
10
2 Difference of bid-value ratio and RNNE ratio N
9
11
2 Difference of bid-value ratio and RNNE ratio N
14
12
4 Bidding function of market size N

• Estimation
• Blind Experiment A 8.66, ? .639

13
6 Concluding remarks

• Bidding in FPA increases with N but not as much
  as the RNNE predicts.
• Prices are above expected prices in the Blind
  experiment for N lt 14. Differences decline
  significantly in N and for N 14 prices are at
  the RNNE.
• The magical number seven plus or minus two,
  Miller 1957 PsychRev.
• Two or three different tones were never
  confused. With four different tones confusions
  are rare, but with five or more tones confusions
  are frequent. With fourteen different tones
  listeners make many mistakes. (p. 344)
• PLEASE SEND ME AN EMAIL TO OBTAIN THE PAPER!
• THANKS!