An analysis of potential Treasury auction techniques

Federal Reserve Bulletin, June, 1992 by Vincent Reinhart

Indeed, the profit from a market squeeze may come by other means. While the issue remains in the cornerer's control during secondary trading, short sellers must borrow the security to make delivery. That transaction is one side of a repurchase agreement in which the owner of the desirable security--the cornerer--lends it to a short seller in return for cash at a preferential bonowing rate. In effect, by creating a demand for the issue, the cornerer can finance his or her position at a below-market bonowing rate.

The when-issued market plays two important roles in cornering strategy. First, early trading allows the market consensus to coalesce quickly and thus provides a usually accurate forecast of the auction price. By aiding in the "price discovery" of the appropriate price on the security to be auctioned, the when-issued market serves in tightening the spread of bids; thus, the cornerer needs to bid only slightly higher than that consensus to be assured awards. Second, a group of thwarted bidders--those who shorted in the when-issued market--are forced to the secondary market to close their positions. Their surprising presence makes the demand schedule less price sensitive, as no substitute exists for the security that they promised to deliver. As a result, as long as they keep their positions open, short sellers will need to borrow the desirable security and thus provide the cornerer favorable financing in the repurchase market.

The successful cornerer makes use of three elements of the current practice:

* When-issued trading creates a core of reliable demanders for the auctioned security (those who sold short).

* The first-price method of allocating awards reduces demand at the auction and makes that demand more price sensitive.

* Sealed bids allow a cornerer to place bids only marginally better than the consensus to win all the awards.

These characteristics of current procedures promise profit in successfully cornering a Treasury auction, although such trades are not without considerable risk. Even slight shifts in the prevailing level of interest rates could more than wipe out the profit from controlling a significant portion of an outstanding issue.

The Potential for Collusion

One dealer with adequate capital and the willingness to be exposed to substantial risk can possibly take advantage in the current market. A harder problem to assess is whether or not an auction's design may entice a group of dealers to conspire in an attempt to corner. The theoretical analysis of the incentives for collusion in auctions proceeds as follows.

Let us suppose that a few dealers, intent on extracting profit from those not in the ting, willfully plan together to purchase all that is sold at an auction. They agree on a price just above the market consensus that is sure to win all the awards. A sealed-bid auction, however, tempts each of the conspirators to move just above the agreed-upon price and to steal awards; as a result, the cartel likely will not hold.9 Hence, on the one hand, incentives in the classic first-price sealed-bid auction are structured so as to make collusion unlikely. On the other hand, in an ascending-price openoutcry auction, such a conniver among conspirators has to show his or her hand, making such manipulation less likely. Even if bidding is secret, the other members of the cartel will know by the price movement that someone has cheated. The cartel will hold.