Introduction to Auction

Summary

The lecture introduces auctions as a form of dynamic pricing. Although fixed prices are convenient and common, the internet has made dynamic pricing practical again. Auctions allow prices to emerge from supply and demand and are used in settings ranging from online marketplaces to spectrum sales.

Auctions are a long-standing market mechanism, but poorly designed auctions can fail. The lecture cites US spectrum auctions and oil-exploration rights as examples of important auction-design problems.

Main Auction Classifications

By the Resources Being Traded

• Single item, single unit: One unique item is sold, such as a painting or a piece of equipment.
• Single item, multiple units: Several units of the same item are bought or sold, such as when a company purchases a large quantity of sugar.
• Multiple items: Different items are offered together. When bidders can make offers on combinations of items, this is a combinatorial auction.
• Multiple items and multiple units: A more complex form of combinatorial auction.
• Multiple attributes: Items may differ in qualities such as warranty or service, as well as price. These attributes need to be assessed together, often using a scoring or utility function.

By Market Structure

• One seller and multiple buyers: A forward auction, in which buyers compete to purchase from a seller.
• One buyer and multiple sellers: A reverse auction, in which sellers compete to supply what the buyer wants. Online systems can allow this competition to take place over multiple rounds, unlike a one-time tender.
• Multiple buyers and multiple sellers: A double auction, such as the matching that takes place in stock markets.

By Bidding Rules

• Ascending-bid (English) auction: Bidders successively raise their offers. The item goes to the highest bidder when no one raises further. The auction needs rules for a minimum starting or reserve price and for permitted bid increments.
• Descending-bid (Dutch) auction: The seller starts with a high price and lowers it until a buyer accepts. The first buyer to accept wins.
• Sealed-bid auction: Bidders submit offers without seeing or revising one another’s bids during the process. Two common payment forms are:
  • First-price: The highest bidder wins and pays their own bid.
  • Second-price: The highest bidder wins but pays the second-highest bid.

By Buyer or Seller Preferences

Preferences can vary with the number of units acquired. For example, a buyer may value each additional unit less as their total quantity increases. Preferences can also depend on non-price attributes, such as quality or warranty. A utility or scoring function can combine these considerations into a common monetary or numerical measure for comparing offers.

By Bid Structure

The information a bidder submits depends on what is being traded:

• For a single item and unit, the bid may consist of just a price.
• For one item in multiple quantities, it may be a price–quantity pair.
• For multiple items, it may list the items desired and the price offered for each item or combination.

By Payment and Winner-Determination Rules

• Under a first-price payment rule, the winner pays their own bid; under a second-price rule, the winner pays the next-highest bid.
• The lecture also mentions all-pay arrangements, in which participants pay to take part even though only the highest bidder receives the item.
• Deciding who wins is straightforward when one buyer or seller can handle the whole transaction. When supply or demand must be split among multiple participants, the problem becomes combinatorial and can be computationally difficult.

By Feedback During Bidding

• No-feedback auctions: Sealed bids are submitted once and cannot be adjusted based on other bids.
• Feedback auctions: In an English auction, bidders can see the current high bid and respond to it.

Practical Considerations for Online Auctions

The lecture highlights four broad concerns:

• Economic design: Choosing an auction format and rules that suit the market.
• Business rules: Setting matters such as reserve prices and whether to use one or multiple suppliers.
• Computation: Determining winners may require solving an optimization problem, particularly in multi-item or multi-source cases.
• Security: Bid information may need to remain confidential during information exchange.

Speakers and Sources Featured

• Speaker: One unnamed lecturer.
• Sources and examples mentioned: eBay and similar online auction sites; the US Federal Communications Commission’s spectrum auctions; US coastal oil-exploration rights; Dutch flower auctions; stock markets; and a Vickrey auction example that the lecturer describes as a failure.