Something to think about ~
For Google, Going Dutch Has Its Rewards And Its Risks
By SAUL HANSELL
Published: May 10, 2004
To the idealistic computer scientists who run Google Inc., the company's unusual plan to auction shares in its public offering is both more efficient and fairer than the traditional method in which investment banks set the offering price.
Doubtless, selling shares to the highest bidders would raise the most money from the offering. It would also eliminate the abuses seen in recent years in which bankers gave shares -- at low offering prices that were sure to rise once trading started -- to favored investors and executives.
But for a company making its stock-market debut, having the highest possible offering price has risks as well, according to experts who have studied auctions in the 20 countries that have tried them. In some cases, auctions caused speculative frenzies, hurting both investors and companies in the long run. In others, the auction structure eliminated the incentive for investors to back new and untested companies, leading to lower prices than the companies expected.
''An auction is just too risky,'' said Ann Sherman, an assistant professor of finance at the University of Notre Dame who is a critic of share auctions. ''There is no telling how many people will show up.''
Google will not comment on anything concerning its planned public offering, including its auction. But it deals with many of the concerns in its prospectus.
Moreover, economists who favor auctions argue that the criticisms pale when compared to the benefits.
''An auction is the fairest and most transparent system and will get the best price for the issuer,'' said Paul Klemperer, an economics professor at Oxford and author of a forthcoming book, ''Auction Theory and Practice.''
Google is going to use a variation on what is known as a Dutch auction, called that because it was created in the early flower markets of the Netherlands to sell multiple identical items. In the classic Dutch auction, a seller indicates how many items are available for sale and sets the minimum bid price.
Bidders indicate the number of items they want to buy and the price per item they are willing to pay. All winning bidders pay the same price per item -- which is the lowest successful bid, called the clearing price. Those who bid above the clearing price, however, earn the right to buy the number of items they want, while those who bid at the clearing price have to divide the remainder.
Google's shares would be sold in a modified Dutch auction because it has reserved the right to set the final sale price, the allocation of shares and other auction terms. It said in its public offering statement that its goal was to eliminate the first day ''pop'' in prices that was built into most initial stock offerings.
Professor Sherman argues that the traditional offering process in which investment bankers set a low initial stock price and allocate shares to certain investors is an appropriate way to encourage investors, especially big institutions, to learn about a new company by reading the prospectus and attending its presentations.
''You can really put in time and effort to do your homework and put in a serious bid,'' Professor Sherman said. But then someone like that, she added, can easily ''get crowded out by 5,000 day traders.''
Indeed, even those who reject traditional underwriting often realize that building in a bit of room into the initial sale price for first-day investors is necessary.
Patrick M. Byrne, the chairman of Overstock.com, an Internet retailer, chose to take his company public by using an auction because he was disgusted with the way shares in initial offerings are typically allocated.
''Any time there are guaranteed profits, there are kickbacks,'' Mr. Byrne says. ''It doesn't matter whether it is a third world customs examiner or a white shoe investment banker.''
But when Mr. Byrne reviewed the bids in the offices of W.R. Hambrecht & Company, the San Francisco-based investment bank that has been a pioneer in conducting auctions, he decided the price had to be set a bit lower than what the market would bear. There was enough investor demand to sell all available shares at $14 each. But Mr. Byrne decided to set the initial price at $13 -- a price that meant investors received less than 60 percent of the shares they asked for.
''We wanted a little tailwind so that you know it will not come out and trade down,'' Mr. Byrne said of the initial offering price. ''We thought it was fair for the people who were stepping up and buying our stock in the I.P.O.''
While that is similar logic to that used by traditional underwriters to justify their methods, Mr. Byrne says there is an important difference: ''The investors deserved a 5 to 10 percent tailwind,'' he said, ''not an 80 percent tailwind.''
While Google reserved the right to sell shares at below the market clearing price, it said it wanted to set the price high enough so that investors would receive at least 80 percent of the shares they bid for.
Another potential problem with Dutch auctions is that investors have an incentive to bid higher than the fair value of a stock so that they can be assured of getting shares to buy. If only a few people bid high, they would still only pay the market clearing price determined by the vast number of presumably more rational investors, since the price would be set by the lowest bid. But if lots of investors take up this strategy, the price would be driven above sustainable levels.
Professor Sherman said that in many big stock auctions in Europe and Asia, the auction process encouraged a speculative frenzy and a destructive backlash. Shares of Singapore Telecom, she said, rose the first day after its 1993 initial offering, but they fell shortly thereafter and took 18 months to rebound above the offering price.
The company, of course, raised money on attractive terms. But Professor Sherman said that the danger to investors and the negative perception caused by a declining share price hurt the company more than it benefited from the additional cash.
''It can be bad for managers to be judged by how they perform after an offering that has a wacko price,'' she said, adding that such auctions have been abandoned in most of the countries that have tried them.
In the United States, the experience of auction offerings through Hambrecht have been quite variable. In 1999, the auction price of Andover.net was $18 a share, but it shot to $67.50 three days later. The next year, shares of Nogatech were auctioned at $12, but three days later it closed at $7.44.
Another factor, called the ''winner's curse,'' can potentially lead to depressed prices after the initial offering. The top bidders may realize that they bid more than what other bidders believe the shares are worth. If the winners start to worry about the share price paid, they may sell shares immediately after the auction, causing a drop in the price. For more thoughtful investors, the fear of the winner's curse could lead them to moderate their bids in advance -- a move that might lead to lower prices in an auction than in a traditional offering in which the price is set by investment bankers.
Google, whose founders are academically inclined, acknowledges most of these auction risks in its prospectus. The document contains a section on the winner's curse and how it could cause Google's shares to decline sharply in the days after the offering. To ward off extreme share price fluctuations, the company has given its underwriters the ability to reject bids that they deem to be speculative (at an unreasonably high price) or manipulative (for an excessive quantity of shares).
Google's version of a Dutch auction would eliminate a primary incentive for placing high bids. Normally, buyers who bid above the actual clearing price are allowed to buy all the shares they ask for and those that bid exactly at the sale price divide the remaining shares, usually receiving fewer than they wanted. Google, by contrast, proposes two methods to allocate shares that would not reward buyers for high bids. (One method would give bidders a set percentage of their orders. The other would impose a maximum number of shares any bidder could buy.)
If Google is successful in moderating the tendency of auctions to encourage speculative bidding, it would get an auction price identical to the price on the first day's trading. But if that is true, there may be no reason for investors to participate.
''For the small investor, there is very little way to game the system,'' said Lawrence M. Ausubel, an auction expert and economics professor at the University of Maryland. ''You should be relatively indifferent about winning or losing the I.P.O. auction, because if Google does what you expect and selects the I.P.O. price to be the true clearing price, you will have the option to buy at essentially the same price the next day.''
Drawing (Illustration by Stuart Goldenberg)
For Google, Going Dutch Has Its Rewards And Its Risks
By SAUL HANSELL
Published: May 10, 2004
To the idealistic computer scientists who run Google Inc., the company's unusual plan to auction shares in its public offering is both more efficient and fairer than the traditional method in which investment banks set the offering price.
Doubtless, selling shares to the highest bidders would raise the most money from the offering. It would also eliminate the abuses seen in recent years in which bankers gave shares -- at low offering prices that were sure to rise once trading started -- to favored investors and executives.
But for a company making its stock-market debut, having the highest possible offering price has risks as well, according to experts who have studied auctions in the 20 countries that have tried them. In some cases, auctions caused speculative frenzies, hurting both investors and companies in the long run. In others, the auction structure eliminated the incentive for investors to back new and untested companies, leading to lower prices than the companies expected.
''An auction is just too risky,'' said Ann Sherman, an assistant professor of finance at the University of Notre Dame who is a critic of share auctions. ''There is no telling how many people will show up.''
Google will not comment on anything concerning its planned public offering, including its auction. But it deals with many of the concerns in its prospectus.
Moreover, economists who favor auctions argue that the criticisms pale when compared to the benefits.
''An auction is the fairest and most transparent system and will get the best price for the issuer,'' said Paul Klemperer, an economics professor at Oxford and author of a forthcoming book, ''Auction Theory and Practice.''
Google is going to use a variation on what is known as a Dutch auction, called that because it was created in the early flower markets of the Netherlands to sell multiple identical items. In the classic Dutch auction, a seller indicates how many items are available for sale and sets the minimum bid price.
Bidders indicate the number of items they want to buy and the price per item they are willing to pay. All winning bidders pay the same price per item -- which is the lowest successful bid, called the clearing price. Those who bid above the clearing price, however, earn the right to buy the number of items they want, while those who bid at the clearing price have to divide the remainder.
Google's shares would be sold in a modified Dutch auction because it has reserved the right to set the final sale price, the allocation of shares and other auction terms. It said in its public offering statement that its goal was to eliminate the first day ''pop'' in prices that was built into most initial stock offerings.
Professor Sherman argues that the traditional offering process in which investment bankers set a low initial stock price and allocate shares to certain investors is an appropriate way to encourage investors, especially big institutions, to learn about a new company by reading the prospectus and attending its presentations.
''You can really put in time and effort to do your homework and put in a serious bid,'' Professor Sherman said. But then someone like that, she added, can easily ''get crowded out by 5,000 day traders.''
Indeed, even those who reject traditional underwriting often realize that building in a bit of room into the initial sale price for first-day investors is necessary.
Patrick M. Byrne, the chairman of Overstock.com, an Internet retailer, chose to take his company public by using an auction because he was disgusted with the way shares in initial offerings are typically allocated.
''Any time there are guaranteed profits, there are kickbacks,'' Mr. Byrne says. ''It doesn't matter whether it is a third world customs examiner or a white shoe investment banker.''
But when Mr. Byrne reviewed the bids in the offices of W.R. Hambrecht & Company, the San Francisco-based investment bank that has been a pioneer in conducting auctions, he decided the price had to be set a bit lower than what the market would bear. There was enough investor demand to sell all available shares at $14 each. But Mr. Byrne decided to set the initial price at $13 -- a price that meant investors received less than 60 percent of the shares they asked for.
''We wanted a little tailwind so that you know it will not come out and trade down,'' Mr. Byrne said of the initial offering price. ''We thought it was fair for the people who were stepping up and buying our stock in the I.P.O.''
While that is similar logic to that used by traditional underwriters to justify their methods, Mr. Byrne says there is an important difference: ''The investors deserved a 5 to 10 percent tailwind,'' he said, ''not an 80 percent tailwind.''
While Google reserved the right to sell shares at below the market clearing price, it said it wanted to set the price high enough so that investors would receive at least 80 percent of the shares they bid for.
Another potential problem with Dutch auctions is that investors have an incentive to bid higher than the fair value of a stock so that they can be assured of getting shares to buy. If only a few people bid high, they would still only pay the market clearing price determined by the vast number of presumably more rational investors, since the price would be set by the lowest bid. But if lots of investors take up this strategy, the price would be driven above sustainable levels.
Professor Sherman said that in many big stock auctions in Europe and Asia, the auction process encouraged a speculative frenzy and a destructive backlash. Shares of Singapore Telecom, she said, rose the first day after its 1993 initial offering, but they fell shortly thereafter and took 18 months to rebound above the offering price.
The company, of course, raised money on attractive terms. But Professor Sherman said that the danger to investors and the negative perception caused by a declining share price hurt the company more than it benefited from the additional cash.
''It can be bad for managers to be judged by how they perform after an offering that has a wacko price,'' she said, adding that such auctions have been abandoned in most of the countries that have tried them.
In the United States, the experience of auction offerings through Hambrecht have been quite variable. In 1999, the auction price of Andover.net was $18 a share, but it shot to $67.50 three days later. The next year, shares of Nogatech were auctioned at $12, but three days later it closed at $7.44.
Another factor, called the ''winner's curse,'' can potentially lead to depressed prices after the initial offering. The top bidders may realize that they bid more than what other bidders believe the shares are worth. If the winners start to worry about the share price paid, they may sell shares immediately after the auction, causing a drop in the price. For more thoughtful investors, the fear of the winner's curse could lead them to moderate their bids in advance -- a move that might lead to lower prices in an auction than in a traditional offering in which the price is set by investment bankers.
Google, whose founders are academically inclined, acknowledges most of these auction risks in its prospectus. The document contains a section on the winner's curse and how it could cause Google's shares to decline sharply in the days after the offering. To ward off extreme share price fluctuations, the company has given its underwriters the ability to reject bids that they deem to be speculative (at an unreasonably high price) or manipulative (for an excessive quantity of shares).
Google's version of a Dutch auction would eliminate a primary incentive for placing high bids. Normally, buyers who bid above the actual clearing price are allowed to buy all the shares they ask for and those that bid exactly at the sale price divide the remaining shares, usually receiving fewer than they wanted. Google, by contrast, proposes two methods to allocate shares that would not reward buyers for high bids. (One method would give bidders a set percentage of their orders. The other would impose a maximum number of shares any bidder could buy.)
If Google is successful in moderating the tendency of auctions to encourage speculative bidding, it would get an auction price identical to the price on the first day's trading. But if that is true, there may be no reason for investors to participate.
''For the small investor, there is very little way to game the system,'' said Lawrence M. Ausubel, an auction expert and economics professor at the University of Maryland. ''You should be relatively indifferent about winning or losing the I.P.O. auction, because if Google does what you expect and selects the I.P.O. price to be the true clearing price, you will have the option to buy at essentially the same price the next day.''
Drawing (Illustration by Stuart Goldenberg)
