"the final offering price should capture the entire market demand. Therefore, when shares begin trading in the aftermarket on the first day, the price should not fluctuate."

But this prediction relies on some very strong and unrealistic assumptions: 1) that all interested investors bid in the auction; 2) that they all devote time and effort to first figuring out how much they think that the shares are worth (or that the number simply pops into their head out of nowhere); 3) that everyone bids the valuation that they came up with, even if it's not in their interests to do so; and 4) that no one's valuation changes once they see the auction results.

You need all 4 of these assumptions to get the "perfect pricing" prediction, and yet not one of the 4 is likely to be satisfied in a "Dutch" auction. In theory, auctions are extremely unreliable at predicting IPO prices if we make realistic assumptions such as: 1) it takes time and effort to come up with a good estimate of the value of the shares; and 2) people are free to decide for themselves whether or not to enter, and make that decision based on their own best interests, as they see them.

Yes, auctions prevent underwriters from abusing their discretion, but only in an extreme way, by eliminating discretion entirely - the classic 'throwing out the baby with the bathwater'. Issuers care about risk, they want to attract serious long term investors, and they have many other goals that aren't best served by simply tossing the shares on the market and keeping their fingers crossed.

Past practice in the US has been far from perfect, and there were some extremes during the bubble period. But forming policy based on a few isolated instances under extreme circumstances is foolish, and turning to a method (IPO auctions) that has been well-tested and has consistently failed is even less prudent.

Ann, thanks for pointing out fallacy that auction IPOs should open flat. I'm not sure why everyone thinks they should. You can add #5: the company will price it shares at a small discount to eliminate any downward pressure.

pb - Thanks for adding to the list. There might be more conditions we haven't thought of.

I think that the intuition about auctions was developed in the private values setting, where each person has their own, fixed, personal value for the object. So, some people really like the painting, some don't. They just look at it and they know if they like it or not. If you get everyone to bid, then the person that likes the painting the most bids the highest. Even after she sees that she has bid more than everyone else, she still likes the painting just as much, because her bid was based on her own personal valuation.

With IPOs, you have huge amounts of uncertainty, and it takes work to come up with a good estimate. But, in a "Dutch" auction, there's also a short cut - bid high, and figure that someone else will do the hard work of actually valuing it. If you bid really high, you're first in line, but you only pay the clearing price or less. If you're the only one with the brilliant idea of bidding high, everything is fine. But with tens of millions of potential bidders, if too many others get the same bright idea, you're all in a lot of trouble.

The reason many countries (for instance, Singapore, Argentina and Turkey) stopped using IPO auctions was because of one or two issues that were spectacularly overpriced, from too many people taking the shortcut at once. It can get very messy. That's why Google warned against the free rider problem (although they mistakenly labelled it the winner's curse - the winner's curse is a separate, much smaller problem).

Anyway, that's probably more than you wanted to know. If you're not put off by academic papers, I have one posted on ssrn that discusses and models much of this (http://ssrn.com/abstract=276124). It also discusses how IPO auctions are different from government debt auctions.

It's a shame to see the debate over the problems with the US IPO method get sidetracked on this inefficient alternative. I wish that Google had tried something truly innovative, rather than an old, failed method.

Great comments. First, I think you've helped me clarify part of what I was talking about in the Catch-22. Perhaps it is not true to say that a good IPO auction must result in capturing the entire market demand and therefore will produce a share price that will not fluctuate. However, when that does happen, commentators call it a failure. Why is that a failure?

I think one point that needs to be clarified is what is a "successful" IPO? Is it an IPO that pops the first day, which some people seem to think. Is it an IPO that places the shares in the hands of long-term investors, an argument I think that you have made? Is it an IPO whose share price closely approximates the fundamental value of the share's percentage in the company, whatever that is? Is it an IPO in which all investors have the same opportunity to participate in the offering?

I do believe that the bookbuilding system is inherently flawed, but I don't see any debate over that. The IPO Advisory Committee, the NASD, the NYSE, and the SEC had the opportunity to change the flaws in the system in the past few years, and they chose not to. If the bookbuilding system is not going to be improved after the 1999-2000 Boom, then it never will be. The "debate" is not being sidetracked by the IPO auction; the debate never got started because the people that benefit from the bookbuilding method are the same people that regulate the system.

I'm also not sure that I would call IPO auctions an "old, failed method" when they have only been legally possible in the US for five years. In addition, auction variants are currently used in countries such as France, Germany, and Israel.

Ann, you commented last night to an earlier post on flipping by institutional investors. At a glance, B. Boehmer, E. Boehmer, and Fishe (2002) cite that in their sample, using the DTC IPO tracking system, that institutions receiving allocations from the lead underwriter sell 25% of their allocation in the first two days. In contrast, retail investors who receive allocations from the lead underwriter sell 11%. As you say, the US market is not transparent, so this data is hard to find. Other authors have used block sales in the first two days as a proxy for institutional sales and come to varying estimates (20% of allocations flipped to 60% of allocations flipped).

I am very interested in this argument that issuers want their shares in the hands of long-term investors. How do we know this? Do issuers actually care who holds their shares? I'm not sure why they would. Some economists have attempted to make connections between block holders of shares and monitoring capabilities, hypothesizing that the more long-term, institutional holders that a company has, the more efficient management will be because the holders have an incentive to monitor. I don't believe that these hypothesizes have been conclusive as to anything, but more importantly, I'm not sure if the issuer wants holders to have that incentive. If the founders are acting in their self-interest, they may want dispersed ownership instead.

Of course, greater institutional ownership may get you positive analyst coverage.

Just an addition -- I found another source in my stack that I wanted to mention. Reena Aggarwal (2003) has written "It has generally been argued that large proportions of an IPO are allocated to institutional investors, the so-called "strong hands," because they are long-term investors and will not flip IPOs in the aftermarket. However, institutions are found to consistently flip a much larger percentage of the shares allocated to them than do retail customers." In Aggarwal's sample set, "[t]hey flip 10.62% (median of 7.22%), 26.40% (median of 22.87%), and 32.07% (median of 25.94%) of the shares allocated to them in IPOs priced below, within, and above the filing range, respectively."

Issuers care about their stock price, which depends on who is willing to hold their stock.

What I think is missing from most of the analysis of IPOs is the fact that it's hard to figure out how much these shares are worth, whether in the IPO or while trading on the secondary market. It's takes work, and there's no reason why any one investor feels compelled to do that work for a particular stock. My view of the current system is that investors are being paid to show up at the road show and give the offering serious consideration. Most IPOs simply aren't important enough to attract much attention without underpricing. Attention and efficient valuation is not inevitable - sometimes one has to pay for services, because people's time is valuable.

A serious concern that companies have is that, after all the time and expense of an IPO, they'll end up in "the Orphanage". Orphans are public companies that don't have analyst coverage, aren't followed by investors, and hence have no ability to do follow-on offerings when they need to. They get the worst of both worlds - all the expense of ongoing reporting requirements (worse now, thanks to Sarbanes-Oxley) but a dead stock and no true market access. That's why they care about attracting serious, long term investors.

Companies also care about their stock price even when they're not planning to issue securities, because their stock price is used as a benchmark in many different ways. A stagnant stock price hurts employee morale, makes customers and suppliers wonder about the long term health of the company and puts the company in a worse bargaining position in negotiations for new projects or additional funding of any type.

If the stock price plunges low enough, eventually the bottom feeders will probably discover it and it may go back to a somewhat more reasonable level. But in the meantime, the company is in a difficult position. Share prices aren't inevitable once the stock starts trading - companies are very aware that they may be overlooked.

Under our current system, institutional investors show up for the road show even if the company is too small or new for them to be interested otherwise, because the underwriter pays them to show up through underpricing. And analysts cover the stock because they have to. If you ask analysts, they'll tell you that they hate covering IPOs, because new companies are risky and harder to cover, they're less high-profile, and there's more potential that the analyst will miss something and be embarrassed.

Analysts cover IPOs for only one of two reasons - because they have to, since they're connected to the underwriter handling the offering, or because there's substantial investor demand. For small, insignificant companies that are at risk of ending up as orphans, the only way to guarantee serious consideration is through the underwriter.

You seem to allow the possibility that issuers might want analyst coverage but not that they care about investors. Why would they value analyst coverage as an end in itself? Analysts simply provide economies of scale in information collection, and they choose which companies to cover based on "market demand". Companies want analyst coverage because it lowers the marginal cost of information for investors, thus attracting more of them.

With an auction, no one has a strong, guaranteed incentive to put a lot of time and effort into valuing the stock, since they all know that they could easily be crowded out. There will still be some evaluation, as I show in my model, but auctions are extremely risky, and the risk of wasting one's time means that bidders won't enter the auction unless they expect to be compensated for their effort and added risk.

I don't know of any legal restrictions that prevented the use of IPO auctions in the US in the past. WR Hambrecht didn't get any laws changed, they just started offering auctions. And I called the auction method old because the UK began doing them more than 20 years ago. Many European countries had already tried and abandoned the method more than a decade before WR Hambrecht pioneered it.

Germany had two IPO auctions. I think that they were both in 1999 or 2000, during the bubble, but they might have been earlier. France still has sporadic use, particularly on the unregulated over-the-counter market, so it's too early to say that IPO auctions have vanished entirely there, but their use has decreased dramatically. As for Israel, the law forcing all issuers to use auctions for 10 years expired in Dec., 2003. I think that there was one debt IPO auction a month or two after expiration of the law, and that two debt IPOs since then used the traditional method that had been banned for a decade. Legislation allowing book building is pending but has been held up.

It's not likely that there will be any more Israeli IPO auctions, now that issuers are allowed to choose, but of course any of this could change anywhere in the world. Overall, though, more than 20 countries have tried IPO auctions, and what's left is sporadic use in the US, France and possibly Israel or Taiwan.

I read your paper and agree with some of your concerns about book building (which is why I have a paper, on ssrn, that proposes changes to the system - http://ssrn.com/abstract=648322). But what I didn't find in your paper was specific discussion of why you think that auctions should be used. Given the many possibilities, ruling out the status quo says nothing about any one particular alternative. What advantages are there to auctions? In terms of IPO methods, auctions have been a distant third choice among issuers worldwide. Why not at least consider #2, before we turn to the one consistent loser?

On flipping, I was thinking of Reena Aggarwal's work but had in mind her paper with Puri and Prabhala. The main point of her earlier paper that you mentioned is that flipping levels of initial investors are lower than people had thought, since many people looked at the total trades the first day and assumed that it all represented sales by initial investors.

I suspect our disagreement on this is over differing interpretations of the optimal level of flipping. The numbers for institutional flipping don't look all that high to me. After all, issuers don't want zero flipping the first day - how could the shares trade? I remember seeing an article once about a French IPO, saying that there had been no trades on the first day of a stock's 'trading', since none of the buyers from the IPO wanted to sell. Having too few flippers can artificially restrict supply, possibly causing a first day pop that's unsustainable (i.e. leading to volatility).

And we'd expect, using past data, to see less flipping among retail, because it has been penalized and restricted. We can't infer from past data what the flipping levels would be if both groups were unrestricted. Market participants view retail investors as being short term, unreliable and very likely to flip, which is why they're restricted in the beginning period of IPOs. And, by the way, hedge funds are viewed the same way.

Under the current system, the underwriter tries to manage who gets what shares, so that there is just enough but not too much first day flipping. If underwriters had no control over allocation, there might be excessive levels of flipping (as has been reported for IPO auctions in Singapore, Malaysia, etc.).

Underwriters are trying to form an active, liquid market for the shares. They're trying to attract long term investors. Yes, in bubble periods even those long term investors may sell out temporarily - many of us would sell our house, even if we really liked it, if someone offered us 5 times what it was worth. But once those stable institutional investors have gotten to know the company, they're more likely to step in later and buy back the stock after the "irrationally exhuberant" have dropped it.

Pricing decisions weren't easy during the bubble period. What is the appropriate offering price for a stock, if certain investors are willing to pay far more than the investment bank thinks that the shares are worth? Should the initial offering price be set at ridiculous levels, virtually guaranteeing a crash later, or should it be set somewhere between the rational price and the fringe price? And if the shares are sold only to the extreme, without ever even being considered by any stable investors, what will happen when it crashes later? There are plenty of examples of IPO shares being auctioned to the highest bidders during hot markets, and the end result wasn't pretty.

Just two clarifications:

1. Before Wit Capital received two no-action letters in 1999 and 2000 from the SEC, no IPO or secondary offerings for debt or equity took place because it was assumed that a true Dutch auction would violate the quiet period rules that no sale or offer for sale take place before the registration statement went effective. Wit Capital was the first I-bank to do IPOs, but they have been subsumed by a traditional I-bank -- perhaps Merrill Lynch, but I don't recall offhand.

2. I understand that an issuer cares about the "stock price" and that an issuer cares about "investors," but I want to know why an issuer would specifically care about whether the investor was a short-term investor or a long-term investor. If you want an "active, liquid market," you have to have short-term investors. If you want the stock price to go up, somebody's going to have to sell, right? I feel like I must be missing something -- that you're saying that long-term investors have a positive effect on the share price. How? (Besides generating analyst interest.)

I do understand about this Orphanage that you describe, but I want to know why that happens. Is it because no large institutional investors hold the stock? Is it because the issuer has no relationship with an I-bank to generate analyst coverage? If it is the second reason, then that seems like an inefficiency in the market for a good company to go untouted because of a market failure for stock research.

What are you touting as #2, fixed price offerings? I am not being a snot; I'm truly interested.

Sorry, I just realized that I never got back to this. In case you're still interested -

Yes, the second most successful method worldwide has been fixed price public offers. In countries such as Singapore, Taiwan and Turkey, fixed price was the established IPO method, and then auctions were introduced (and book building wasn't around, either by regulation or because they just hadn't gotten around to trying it). Auctions were most popular their first year, but became less popular (less likely to be chosen by issuers) over time. Eventually, usually in 2 or 3 years, issuers went back to public offers. Then, years later, book building was introduced and became popular.

Companies want a mix of short and long term investors - short term for liquidity, long term because the company can't maintain a good stock price unless someone actively follows and believes in it long term. With book building, the underwriter can try to balance out short and long term investors. With auctions, no one has control over anything - all issuers can do is to keep their fingers crossed.

The Orphanage could be viewed as a "market failure" relative to a perfect, frictionless, full information world. But costs of evaluation are a fact of life that have to be considered in finding the optimal policy. The Orphanage occurs because it takes time and effort to separate out the underpriced from the overpriced firms. How many small investors with 'day jobs' spend 15 hours a day doing fundamental research on small, obscure companies? Investment bank analysts cover companies because there's a demand from someone - either from the company, with whom the investment bank has a relationship, or from investors. If a company is small and overlooked, analysts won't bother to cover them because there's no investor demand for the reports, and fund managers won't bother because they couldn't trade in the stock anyway, since it's too small and illiquid.

If a company's stock price falls to ridiculous levels, then the bottom feeders should find it. But that only works for extreme cases, and it takes time and is a painful process for the company.

How can a company generate coverage if it believes that it's good but overlooked? It can try to make it worthwhile for an investment bank to cover it, by developing a relationship. Thus, these nasty little inside relationships are a potential way around a market failure caused by the cost of evaluation.

I've worked on these issues for years and have looked at many different systems. I agree with you that there are big potential problems and conflicts with book building, and I have a paper with a co-author proposing changes to the system. But, before we change the system, we need to recognize that there are some advantages to the way it currently operates.

I think that the key point that gets overlooked in most analysis is that it's hard to figure out how much a company is worth, especially during an IPO but also during aftermarket trading. There's a lot of uncertainty and risk, and it's just not optimal to spend huge amounts of time evaluating every single company. Much analysis of IPO auctions assumes that people just wake up one morning knowing how much to bid. If that were true, auctions would work very well.

With book building, the underwriter has huge discretion in allocating shares. Discretion means a potential for abuse, but it also allows the underwriter to provide more services and reduce risk for both issuers and investors, so removing that discretion doesn't automatically result in a superior outcome. After all, putting decisions in the hands of judges and juries opens up the possibility that they will misuse their power. Does that mean that we should prevent innocent people from being convicted by never convicting anyone?

Getting back to the choice between auctions and fixed price public offers -
both methods take away the underwriter's discretion in allocating shares. The key difference is whether the offering price is set by bidders with, at best, mixed incentives and limited information, or by the underwriter and issuer. The many, many problems with IPO auctions in practice, and the pretty consistent preference among issuers for fixed price over auctions (once they had experience with auctions), imply that auctions haven't been the best price-setting mechanism. Thus, if you want to take away the underwriter's discretion over allocation, experience indicates that this might best be done through fixed price rather than auctions.

Auctionindex is an auction site search directory for users to find auction venues from which to buy and sell. Auctionindex was created to aid auction sites that would never be found on other search engines.

Offering free and pay per click listings to site owners and is open to all auction sites. Auctionindex will help to increase website traffic and exposure for all sites listed.

Auctionindex is an auction site search directory for users to find auction venues from which to buy and sell. Auctionindex was created to aid auction sites that would never be found on other search engines.

Offering free and pay per click listings to site owners and is open to all auction sites. Auctionindex will help to increase website traffic and exposure for all sites listed.