An analysis of potential Treasury auction techniques

Federal Reserve Bulletin, June, 1992 by Vincent Reinhart

By this theoretical argument, one might surmise that the current first-price sealed-bid auction protects, at least, against the willful joining of dealers to exploit the Treasury and other dealers. Unfortunately, a gap exists between models and reality, as the rule limiting awards to 35 percent of the issue paradoxically turns incentives back toward collusion. If a conniver plays within the lines of the 35 percent rule, he or she will not win enough securities at the auction to control the secondary market. Consequently, tough enforcement of quantity limits more strongly binds conspirators together.

More to the point, theoretical analyses of collusion assume that a small number of colluding parties share information, an assumption that ignores the multiple arenas in which dealers compete. Dealers will not cooperate in auctions if such cooperation jeopardizes their trading in the secondary market. Given the large number of participants and the apparent mistrust among dealers, auction format is unlikely to bring them together.10 Thus, from the standpoint of public policy, the chief risk seems to lie in the manipulative actions of a single dealer, the rogue with capital, which threaten the integrity of the market.

A CLOSER LOOK AT A POPULAR PROPOSAL FOR REFORM

The abuses of the auction rules last summer rekindled enthusiasm for a simple alternative, the second-price sealed-bid auction, to the current discriminatory pricing practice. Proponents argue that awarding securities at a uniform price rather than at the bid prices would end cornering attempts by eliminating the profit potential in market manipulation. And in a way that sounds contradictory, they argue that total revenue would increase by the surrender of the ability to discriminate across bids.

The Consequences for Revenue

The algebra required to calculate an optimal bidding plan in a multiple-unit auction quickly becomes intractable. No analyst yet has worked through the strategic implications of a large core of bidders carving up a block of securities. The logic of the single-unit case, however, suggests that the extent of bid shading can be extreme. In a firstprice auction of multiple units, a strategic bidder does not have to beat the participant with the next highest valuation to win but must better only the middle of the pack of bidders.

If one steps away from the explicit modeling of bidder behavior, the implications for revenue can be spelled out in terms of shifts in the demand schedule for the auctioned security.[1] As shown in diagram 7 (which repeats the middle panel of the three-figured determination of market prices), part of the Treasury's total revenue results from its charging winners the price that they bid, which for its current practice is measured by the area under the demand schedule labeled "First price." That price discrimination, however, discourages some demand, as investors shade their bids for fear of the winner's curse. Adopting a second-price system turns part of that surplus back to the bidders, shifting out the demand schedule to the position labeled "Second price." Under a first-price scheme, the Treasury would have to work down the left demand schedule and award securities at lower prices to place the total issue (marked by the vertical dashed line). Under the second-price scheme, one price, depicted by the horizontal line drawn to intersect the right demand schedule at the issuance size, exhausts the issue. The consequences for revenue depend on whether or not the loss from the inability to price discriminate (left triangle) is greater than the gain from added demand (right triangle).