I recommend reading this letter from Darrell Issa and friends to the SEC about reforming IPO practices, it is a nice mix of boring and thought-provoking and subtly crazy.
The basic intuition here is (1) it is weird and un-American that IPO prices are set by the judgment and good graces of investment banks rather than "the market," and (2) it is weird and suspicious that on average those prices are mostly "too low" compared to the post-IPO market price. This creates a bit of cognitive dissonance in the letter, which was of course sparked by the Facebook IPO: Facebook's offering price was it is safe to say "too high," and so the lawmakers express some pious concern about whether investors were ripped off, but then spend most of their letter lamenting how "undefinedmposing non-market-based charges, by under-pricing IPOs, places a direct drag on economic growth - it is a tax on the issuer of capital akin to hurting the goose that lays the golden eggs."
Painful though that metaphor is, the thought is maybe not too dissonant: non-market prices mean that someone gets ripped off; that's usually issuers but every so often it's investors. Also the compositions of the ripper and rippee investor groups are different; the lawmakers cite evidence that institutions get allocated more hot IPOs while individuals get allocated more dogshit IPOs so I guess "non-market-based" pricing does in fact help the smart and hurt the dumb. I submit to you that so does, y'know, market-based pricing but whatever.
The congressional letter argues interestingly that the IPO underpricing is caused by litigation risk: if an IPO craters, everyone will sue claiming that the company lied about everything, even if they didn't, while if the IPO does well, no one will sue even if the company lied about everything, because money soothes being lied to. So banks con issuers into underpricing their IPOs to avoid litigation risk, and the banks use this underpricing to pay off their favored institutional clients. For this proposition Issa et al. cite this somewhat apologetic 2002 paper proposing that litigation risk may in fact lead to IPO underpricing, though the main cause is of course asymmetric information: investors who don't know the company are buying from insiders who do, and so charge a premium to insure against the likelihood that the sellers are knowledgeable top-tickers.
So one of their recommendations is to make it harder to sue companies for misleading investors, which is what I thought the JOBS Act was supposed to do but there's always room for improvement. Another, stranger recommendation though is for the SEC to consider requiring all IPOs to be Dutch auctions, where anyone can submit bids and the deal prices at whatever the clearing bid is, because that "reflect[s] free market ideals and provide[s] ordinary investors with a unique opportunity to participate alongside institutions."
Sure whatever.* It's worth noting though that there is actually a free market now: nobody is particularly prevented from doing Dutch auction IPOs, and some companies do. Not that many, though, because they don't go that great - Google, the most famous one, whiffed on both size and price, and it was Google. To a first approximation, issuers seem to think of IPOs as more like the M&A market, where contacts and negotiation are important to achieving the best price and where sometimes auctions are restricted to convince the best buyers to pay more, than like say the Treasury market, which is liquid enough that you can just throw an ad on Craigslist saying how much you have for sale and the best possible bids will roll right in. And to a first approximation I'd guess that issuers and their bankers are the ones who know the most about what marketing and pricing strategies will get the best all-around result for companies and shareholders, in terms of price, size, and future access to capital. And if they thought Dutch auctions worked best, they'd use them more.
* So: if you are a retail investor a book-built IPO hurts you because (1) if there's a lot of demand from institutions you get zeroed and (2) if there's no demand from institutions you get filled and the price probably drops. Fine. But in a Dutch auction either (1) you bid less than big institutional investors and get zeroed or (2) you bid more than big institutional investors and get filled and the price probably drops. Roughly speaking, those are kind of the same.