Reverse Auction Bidding: A statistical review of the first case study
MetadataShow full item record
It was in 2004 that the first case study was done by on the ongoing Reverse Auction Bidding at Texas A&M University. This long-term study has developed from a single case study, completed by van Vleet, to a series of case studies, now combined with personality testing of all participants. van Vleet developed a Microsoft Access database system and Active Server Pages web based user interface for the study. The first case study involved five participants with no prior experience in Reverse Auction Bidding. A study with five participants is considered competitive in accordance with the standard economic Herfindahl Index. van Vleet, concluded that the results showed a level of co-operation in the bidding game between the nominal competitors. In 2010 John Nichols coined the term "tacit collusion" to identify this apparent behavioural pattern observed in the bidding. A significant element of the studies from 2005 to 2009 has been to investigate the "tacit collusion" behaviour. Tacit collusion is not considered an illegal economic behaviour. In 2006 Seth Gregory encountered significant problems with a study involving ten participants using the Access database, as a result of Access' limitations on the number of connections. Gregory's study was migrated to a Microsoft SQL database that was developed by Wellington (2006) and which overcame the limitations. SQL database systems can generate a significant quantity of data which create a computer science problem, now commonly termed 'Data rich - analysis poor'. This study is the first in a series of studies to undertake a detailed statistical study of the early case studies to provide a set of algorithms for development of SQL queries for automated real-time data analysis of future Reverse Auction Bidding case studies. This study showed that a fifth order polynomial fit the contract time compared to the job number. Analysis of the number of bids per minute for the fifteen minutes of bid time showed a log-polynomial equation which provided a reasonable fit to the data. Two sub-games were postulated to describe the operational aspects of the auction. The first game, termed the - game, is between the players with the objective of maximizing average return and the second game, termed the - game, has the objective of average cost minimization for the purchasers and maximization of revenue for the seller group. In conclusion, Reverse Auction Bidding systems are not bid shopping, but the tenet that the purchaser will reduce costs in this type of system compared to the traditional closed bid system is not confirmed with van Vleet's data and any careful consideration of the results of canny players in the 'game suggests higher than average returns for some bidders. The results show a number of patterns in the data that warrant further study, particularly the characteristics of the canny players in the alpha game suggests higher than average returns for some bidders. The results show a number of patterns in the data that warrant further study, particularly the characteristics of the canny players.
Guhya, Dhaval C. (2010). Reverse Auction Bidding: A statistical review of the first case study. Master's thesis, Texas A&M University. Available electronically from