Ch. 3: Auctions in Corporate Finance 131
separation in some cases; in others, cash payments or large non-pecuniary bankruptcy
costs are needed to achieve separation (so that the highest value bidder can be iden-
tified). However, relative to cash bids, bids that involve debt or equity distort ex-post
effort choices. Bids that involve high debt and low equity rank higher because they dis-
tort effort less. Convertibles can work better as they give the seller the option to affect
the ex-post capital structure of the target firm. The model thus is capable of explaining
why debt and convertibles are often part of reorganization plans, and why companies
often end up more highly levered than when they were distressed (Gilson (1997)).
Hansen and Thomas (1998)apply the model ofFrench and McCormick (1984)to
argue that uncertainty surrounding a bankrupt firm’s assets can cause auction prices to
be low. Using the French and McCormick model, with free entry of bidders, the auction
price will beN∗Cless than true value, whereN∗is the equilibrium number of bidders
andCis the pre-bid cost of entry (which they model as an information acquisition cost).
Theoretically, then, the question is whether a court, by having to only obtain one (good)
evaluation of the firm’s assets, can hold costs belowN∗C. They argue that the greater
the uncertainty surrounding a firm’s assets, the worse an auction will perform. By way
of example,Reece (1978)shows that with high uncertainty, a common-value auction
yields a price only 70% of true value.
4.8. Share repurchases
Companies frequently buy back their shares through either fixed-price tender offers or
Dutch auction mechanisms. In a Dutch auction repurchase, a company determines a
quantity of shares to buy back and asks shareholders to submit bids specifying a price
and quantity of shares that they are willing to sell. The bids are ordered according to
price (low to high), creating a supply curve. As the Securities and Exchange Commis-
sion prohibits price discrimination, a uniform price is set corresponding to the lowest
price that enables the firm to buy the pre-determined number of shares.
While there has been little formal modeling of the Dutch auction repurchase process
itself (possibly because no real auction-theoretic issues are present) there is considerable
empirical study, and their effects relative to fixed-price offers has been studied in a
more traditional corporate finance setting.Bagwell (1992)studies 32 Dutch auction
repurchases between 1988 and 1991. In one transaction, the highest bid was 14% above
the pre-announcement market price, while the lowest bid was only 2% above. Such
disparities in bids are documented for the entire sample, showing that the firms did face
upward-sloping supply curves for their shares, contrary to naive ideas of a perfect capital
market. Bagwell mentions several possible explanations, including differences in private
valuations (for example, because of capital gains tax lock-ins), asymmetric information
about a common value as inMilgrom and Weber (1982a, 1982b), or differences in
opinion (Miller (1977)). While tax considerations could play a large role, it is certainly
not a stretch to assume that shareholders will have different information on the value
of a company (even though they share the public information embedded in the current
price).