Will Google's Dutch auction go well?

It sounds great: cut out the investment banking fees and just offer a straight Dutch auction on the stock. After all, aren't auctions the perfect market institution?

Co-blogger Alex thinks that the investment banks have had a comeuppance due for a long time; he may well be right.

Under standard practice, the underwriters give underpriced shares to favored investors and executives. The value of those shares rises on opening day. The insiders are happy but the company has left money on the table. In extreme and indeed pathological cases the discount can be as high as eighty percent. So why have companies tolerated this practice for so long?

Under one apologetic view, the kickbacks, underpriced shares, and payola are necessary. Someone has to produce reputation for the stock. The investment bank is paid to do this. The underwriter, in turn, gives insiders good deals to get them to boost the stock. If you own some shares you will do your best to talk up the issue. The efficient markets hypothesis? Well, it may be true at the margin, but how do we get to this margin in the first place?

I heard another point from one industry insider. Investors feel better about a stock if it goes up on day one. For the long-run good of the stock, it is important to have a price rise in the beginning. If investors sour on the stock in its early days, it may never recover its reputation.

Other skeptics wonder how the results of the auction can be predicted? How many people will show up with bids? What if we gave an auction and nobody came? Other worriers fear the temptation for untutored investors to bid too high at first, pushing the shares to unsustainable levels. After all, no single investor will have the final price much with his or her bid.

There is yet another fear. If the auction is fair, the stock will sell at roughly the same price on day one and day two. So if there is some uncertainty surrounding the initial auction, why not just hold off your buying until day two? But then how do liquid markets get established in the first place? How can you get concentrated buying interest on day one, but without violating either fairness or the efficiency of markets?

The resolution: ...will have to wait for the facts and thus the actual auction. But my suspicion is the following. Some percentage of the original underpricing, but by no means all, is in fact a legitimate return to the investment banks. I thus worry that Google will not see strong demand on day one. On top of that, there is a puzzle. Unless you think all of the initial share underpricing is an legitimate fee for services rendered, why have markets tolerated this practice for so long?

By the way, David Levy informs me that used book dealer ALibris will try a share auction as well. For whatever reasons, except for Google, only small companies have shown an interest in these alternative institutions. France uses such auctions more commonly; it seems that the first-day price run-up is smaller but still present. On one hand, these other examples suggest that the auctions are a viable institution. On the other hand, it makes you wonder why the practice is not used more often.

Addendum: Here is a very good piece on the auction of Salon.com stock.