Auction theory and policy

1
Auction theory and policy

• Why auction theory matters for governments
• Europe and the 3G auctionsThe UK and the
  Netherlands
• Auctions and the Victorian Government

2
1 Why auction theory matters for government

• Governments
• Buy goods (and services)
• Sell goods
• Auctions overcome information problems (adverse
  selection)
• Auctions are transparent

3
1 Why auction theory matters for government

• Auction design impacts on behaviour, which in
  turn impacts on the outcome of the auction
• Auctions can be tailored to reflect
• The governments objectives
• The good/service being auctioned
• Single units or bundles of units? Do
  complementarities exist between units?
• The nature of the market
• Competitive industry? (Market power a problem?)
• Collusion likely to be a problem?

4
2 Europe and the 3G auctions

• What is a 3G license?
• The policy issue
• In 2000 six Governments across Europe auctioned
  off 3G licenses.
• 3G new technology governments had no capacity to
  estimate the value of licenses.
• 3G licenses had a common value component
• Each country designed its own auction.
• Each market had different market conditions
  (number and size of incumbents and potential
  market entrants).
• Countries made different numbers/bundles of
  licenses available for purchase.

5
2.1 The UK auctions

• The UK market conditions
• 4 incumbent companies with market advantages
• Several potential market entrants
• Government goals
• Assign licenses efficiently
• Encourage competition
• To realise the full economic value of licenses
• Limit of one license per bidder (to maintain a
  competitive industry)
• The Government hired auction theorists to design
  the auction and tested the design in experimental
  labs.

6
2.1 The UK auctions cont

• The original plan was to sell 4 licenses.
• Because there were 4 incumbents there was concern
  that an ascending auction (English Auction)
• Would discourage new entrants, as dominant
  companies can always out-bid weaker firms
• May lead to collusion
• On the other hand there was concern that a
  sealed-bid auction may lead to an inefficient
  outcome, as dominant firms may under-bid in an
  attempt to capture large surpluses, and therefore
  be outbid by weaker firms

7
2.1 The UK auctions cont

• The Government therefore planned to run an
  Anglo-Dutch auction an ascending auction would
  run until 5 contestants remained, at which point
  these companies would submit sealed bids at or
  exceeding the final price.
• This format performed well in experimental
  testing.

8
2.1 The UK auctions cont

• At this point, a fifth license became available
  for sale.
• As a result the government elected to use a
  simple ascending (English) auction - the
  restriction of one license per company and the
  availability of a fifth license made it possible
  for a new market entrant to purchase a license.
• 9 new market entrants participated ( 4
  incumbents)
• The UK 3G auction raised 39billion
• This auction format gave the government a good
  understanding of the value the market placed on
  3G licenses.

9
2.1 The UK auctions cont

• What can be learned from the UK 3G auction?
• There is no such thing as a good auction, just
  a good auction for the particular task.
• UK Government hired auction theorists and tested
  their auction design in the lab to make sure that
  their design was robust and achieved their goals.

10
2.2 The Netherlands auctions

• The Dutch market conditions
• 5 incumbent companies with market advantages
  (versus 4 in the UK)
• 5 licenses for purchase (as per the UK)
• The Dutch Government used the same design as the
  UK Government an ascending auction.

11
2.2 The Netherlands auctions cont...

• Recognising that the 5 incumbent companies would
  be able to out-bid them, smaller companies chose
  to make deals with the incumbents rather than
  losing at auction.
• Only 1 non-incumbent participated ( 5
  incumbents)
• The Netherlands auction raised 3 billion (versus
  39 billion in the UK).
• The auction did not reveal to the Dutch
  Government the value the market placed on 3G
  licenses.

12
2.2 The Netherlands auctions cont...

• What can be learned from the Netherlands 3G
  auction?
• The auction format that works in one set of
  circumstances is extremely unlikely to work
  successfully in other circumstances subtle
  variations in the market conditions can give rise
  to big differences between outcomes. Treating
  auction designs like blueprints is extremely
  risky for governments.
• Auctions needs to be tailor-made auction theory
  and experimental testing will substantially
  improve auction outcomes.

13
3 Auctions and the Victorian Government

• BushTender
• The Government wanted to improve native
  vegetation on private land therefore needed to
  procure improvements in native vegetation.
• The Government wanted to engage private
  landholders in contracts to improve their native
  vegetation.
• A specific goal was that contracts should be
  cost-effective.

14
3 Auctions and the Victorian Government

• Problems
• It was difficult to measure improvements in
  native vegetation the good the government
  wanted to purchase was poorly defined
• It was difficult to identify the landholders who
  offered the government good value for money
  (adverse selection)
• Solution
• The Government developed a metric that made it
  possible to measure the quality and quantity of
  vegetation improvements that they were purchasing
  (Habitat Hectare)
• The Government decided to use an auction
  mechanism to allocate contracts for improvements
  in native vegetation.

15
3 Auctions and the Victorian Government

• Design questions
• First price or second price? (Discriminating
  price or uniform price auction)?
• Sealed bid or open auction?
• Single round or multiple rounds?
• How encourage participation?

16
3 Auctions and the Victorian Government

• The piloted BushTender format
• First-price auction (discriminating price
  auction)
• Sealed bid
• Single round
• Encouraged participation through lengthy
  engagement with potential participants and
  information distribution.

17
3 Auctions and the Victorian Government
18
3 Auctions and the Victorian Government
Rare and good value
Common and poor value
19
Auction theory and policy

• Summary
• Auctions are a transparent allocation mechanism
• The design of an auction impacts on participants
  behaviour (bidding), and in turn the outcome of
  the auction
• Well-designed auctions can be used to
• Improve allocative efficiency
• Increase revenue / reduce expenditure
• Gather information
• Auctions must be carefully designed to reflect
  policy objectives and the specific nature of the
  good/s being auctioned and market conditions
• Auction theory and can be used to improve auction
  design, while testing in experimental labs can be
  used to test the robustness of an auction and to
  minimise the risk of unexpected outcomes