Peter Cramton writes: Dear Al,
Here is the longer version of my paper with Larry Ausubel on possible details of a troubled asset reverse auction. It might even be a good topic for your market design class.
"A Troubled Asset Reverse Auction" (with Lawrence M. Ausubel), Working Paper, University of Maryland, September 2008. [Presentation]
A Troubled Asset Reverse Auction
Lawrence M. Ausubel and Peter Cramton
30 September 2008
Summary
The US Treasury has proposed purchasing $700 billion of troubled assets to restore liquidity and solve the current financial crisis, using market mechanisms such as reverse auctions where appropriate. This paper presents a high-level design for a troubled asset reverse auction and discusses the auction design issues. We assume that the key objectives of the auction are to:
· provide a quick and effective means to purchase troubled assets and increase liquidity;
· protect the taxpayer by yielding a price for assets related to their value; and
· offer a transparent rules-based process that minimizes discretion and favoritism.
We propose a two-part approach.
Part 1. Groups of related securities are purchased in simultaneous descending clock auctions. The auctions operate on a security-by-security basis to avoid adverse selection. To assure that the auction for each security is competitive, the demand for each security is capped at the total quantity offered by all but the largest three sellers. Demand bids from private buyers are also allowed. The simultaneous clock auctions protect the taxpayer by yielding a competitive price for each security and allow bidders to manage liquidity constraints and portfolio risk. The resulting price discovery also improves the liquidity of the securities that are not purchased in the auctions.
Part 2. Following Part 1, the remaining quantity is purchased in descending clock auctions in which many securities are pooled together. To minimize adverse selection, reference prices are calculated for each security from a model that includes all of the characteristics of each security including the market information revealed in the security-by-security auctions of Part 1. Bids in the pooled auctions are specified in terms of a percentage of the reference price for each security.
The two parts are complementary. Part 1 quickly adds liquidity and establishes competitive market prices for many securities. Additional assets are then purchased in Part 2, taking full advantage of the market information revealed in Part 1.
Clock auctions have been used successfully in the electricity and natural gas sectors for assets worth tens of billions of dollars over the last seven years. Related approaches have been used to auction emission allowances, radio spectrum, timber rights, and rough diamonds.
The approach is feasible even on an extremely tight timetable. The development of the reference price model and data will require the most time to complete, but fortunately this is not needed until Part 2. The auctions of Part 1 can begin in a matter of weeks.
We limit the scope of the paper to taking the legislation as given and providing an auction design within the requirements of the legislation.
The paper is organized as follows. We begin by explaining the adverse selection problem that arises when different securities are pooled together. Then we provide an overview of the two-part reverse auction plan. Next we develop details of Part 1, the security-by-security auctions. We then present the motivation for the simultaneous descending clock auction. Next we present further details of Part 2, the pooled auctions. Then we explore how stock warrants and senior debt can be usefully incorporated into the purchases of troubled assets. Next we discuss the feasibility of successful implementation on an expedited schedule. We conclude that the design achieves the key objectives and can be implemented with minimal risk in a short period of time.
Wednesday, October 1, 2008
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment