Is Underwriting a Cartel?

Felix Salmon posted a striking graph earlier this week, showing just how tightly US IPOs cluster around a 7% commission rate.
This has puzzled analysts for a while.  Why isn't the market for IPOs more competitive?  Felix calls banking a "cartel", but cartels are actually very hard to maintain without legal backing unless it's very hard to enter your market, or you have a good punishment mechanism for defectors.  Otherwise, one of two things will happen: the excess profits will attract new entrants who undercut you, or the members of the cartel will start cheating.

So why hasn't this happened?  One possibility is that the big players may have a good extralegal way to punish defectors.  The major underwriters also buy and sell a lot of stock, both for clients and for their own account.  And they have big equity research shops.  is it possible that this gives them enough power to keep the other banks in line?

It is possible, but there are reasons to be skeptical, too.  For one thing, that's an awfully big conspiracy.  For another, there has been entry into the market over the last few decades.  Why haven't companies used the new entry to beat down the fees--or shifted to auctions, as Google did in its IPO?  You often hear claims that banks are somehow taking advantage of naive managers, but the venture capitalists behind many of the firms that IPO are anything but naive.  They're sophisticated, just as greedy as the bankers, and they're in it for the long haul.  It would be to their great benefit to drive down the fees on these transactions.

I took a crack at this question a few years ago, and the not-very-satisfactory answer I came up with is the managers of companies going public are just not very price sensitive.  I compared it to Don LaFontaine, the voice-over actor who was said to make as much as a quarter of a million dollars for a few minutes work.

When a studio spends tens of millions of dollars producing a film, and further tens of millions advertising it, cheaping out on a voice-over makes no sense. You could pay a Don LaFontaine successor $300,000 a spot and still eat up just a tiny percentage of the film's overall budget. A bad voice-over would cost you far more than you could hope to save. When you have only one chance to get it right, you tend to open up your wallet and pray. So one-shot deals are very, very expensive--a logic that prevails with weddings, funerals, and college diplomas.

That same logic explains why clients have been willing to pay investment banks lavish fees to do IPOs--and secondary offerings, and bond underwriting, and M&A, and advisory work. The mystery of investment-banking fees is often framed as a matter of banks rooking naive managers, or managers selling out their shareholders in return for a space on Merrill Lynch's private jet. But the venture capitalists behind many of the IPOs aren't neophytes at the mercy of big-city bankers; both they and the firm's managers depend on a strong IPO, and a liquid aftermarket, to allow them to get some of their money back out of the company. If they're tolerating such large fees, there must be a reason.

When the boutique firm WR Hambrecht + Co persuaded Google to use an auction process for its 2004 IPO, there was a lot of talk about the end of traditional investment banking. Five years later, however, firms like Goldman Sachs are as dominant as ever, and auctions are rare. Google could rely on its own brand to sell the stock and create a deep secondary market in which its employees could exercise their stock options. But most companies need a little more help. Although Goldman Sachs may not be a bargain, if you're undertaking a one-shot deal, you may want to pay more to hear the one thing a hungry upstart can't tell you: that the company knows how to handle an offering of size and complexity, because it's done so a bunch of times before.

Of course, that doesn't explain why Americans would be less price sensitive than Europeans.  Some possibilities: American regulations requiring lengthy lock-ups may make the managers and venture investors more dependent on the aftermarket than European investors are, meaning that American customers have a stronger incentive to stay on the good side of the banks.  Cosy European banking relationships might allow bankers to make up for lower commissions in other ways--by, say, systematically underpricing or overpricing the IPO and exploiting the  price differentials.  Or European regulators might better be able to punish collusion.   

But these are "just so" stories; the truth is, I don't know.  I can't shake the suspicion that it has something to do with the difference between our stock cultures: Americans are more likely to own stocks than most Europeans, and our markets and our regulations are set up accordingly.  But I admit I don't know what the mechanism is that connects this fact to higher IPO fees.

Update: Commenter WG Gold says

The thing is that every single deal involves several members of the cartel i.e. the lead underwriter usually cuts other players in on the deal to some extent.

 Nobody's ever been so confident in their own pipeline of incoming business that it seemed worth the risk of being marginalized from everyone else's for whatever period of time was necessary for word to get around about their lower prices. 

And Brian Sierk says

I think that part of the issue is that there is a secondary market at play here-most of the guys I know that are or were in finance want to work at a big bank for at least part of their career, both for the money and for the resume polishing.  The Oligopoly , I think is based on the power of the largest banks to negatively impact an individuals career if he or she starts discounting commissions to the point of causing a market shift.   

 It's also possible that there is a social aspect to it.  When i was flipping houses, I negotiated with my realtor on commissions, and she flat out told me that she wouldn't do it because it would harm relationships with other realtors.  There were lots of things that could be forgiven, but undercutting commissions was not one of them.  

The realtor cartel has lasted a shockingly long time; it's now being undercut, but it took decades.  So perhaps the answer is simply that while cartels aren't stable, there's no guarantee as to when they'll fall apart--and in a business where the customers aren't so price sensitive, and the transactions depend on relationships with other players, they can persist for quite a while.