Mark Abrahamson, Tim Jenkinson, and Howard Jones, of Oxford University, have an utterly compelling paper out proving that there’s collusion among investment banks in the US — it doesn’t matter whether they’re European or American banks — to keep IPO proceeds set at 7%. Using a very high-quality new dataset, they compare US and European IPOs, and get the following result:
This chart just shows IPO fees for deals between $25 million and $100 million (in 2007 dollars). But the pattern is universal:
Between 1998 and 2007, 95.4% of U.S. IPOs between $25m and $100m had gross spreads of exactly 7%. The comparable figure between 1989 and 199… While Chen and Ritter showed virtually no IPOs over $150m with a 7% gross spread, we find that 77% of all offerings between $100 and $250m charge exactly 7%.
European IPO fees do not cluster, and only 1% of offerings raising $25m or more experience gross spreads as high as 7%. Within the $25m-$100m range, fees for European IPOs average just over 4%. Indeed, European IPOs are always cheaper: we find that there is a “3% wedge” between European and U.S. IPOs after controlling for size, issue characteristics, syndicate structure and time or country effects. Fourth, whilst gross spreads are lower for the larger offerings in both regions, our multivariate analysis indicates that fees for the larger U.S. IPOs have tended to increase in our sample period, while European IPOs have been getting cheaper.
The paper runs down a list of possible reasons why US IPOs might be so much more expensive than their European counterparts, and finds none of them convincing; their conclusion — the correct conclusion, I think — is that there’s an implicit cartel in the US, devoted to keeping IPO fees artificially high. (The term of art is “strategic pricing”: although it might be in any bank’s short-term interest to compete on price for any given deal, it’s in all of their long-term interest not to ever do so.)
The cost to issuers of this collusion is huge:
Our best estimates suggest that had these IPOs been conducted at European fee levels the savings to U.S. issuers over the period would have totaled $11.4 billion – or over $1 billion per year.
But are there other reasons why the US might have such high fees? I can think of two possible ones. The first is simply cultural: New York bankers won’t haggle on such fees, while their European counterparts do. But that reason is fundamentally circular: it basically just restates the question rather than providing the answer.
The second possibility is that US IPOs are much less important, when it comes to capital raising, than European IPOs. If US IPOs are used mainly to provide pricing, with most capital raised in follow-ons or convertibles, then the fees associated with the IPOs would be less important than in European IPOs, if those were actually used to provide important operating capital.
Here are two charts I just put together, with data from Thomson Reuters’s SDC Platinum:
If there’s a difference here, it isn’t a big one. European IPOs are maybe a little bit more important for capital raising, but nothing to explain the kind of discrepancies found in the paper. There’s an IPO cartel in the US, and now that Europe has moved to a US-style bookrunning model for IPOs, it’s more obvious and blatant than ever in comparison to Europe. Not that anybody’s going to do anything about it.