The US IPO cartel

By Felix Salmon
June 7, 2011
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%.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

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:

ipo.tiff

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:

USIPOS.png

EuroIPOS.png

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.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
15 comments so far

“Cartel” and “collusion” are quite loaded terms to use. “Rational price equilibrium” would be another way to describe it.

Posted by right | Report as abusive

“Rational price equilibrium” might be appropriate if the fees were driven by cost, which they are not. There is no way that the cost of IPOs is always going to be near 7%, that is just an arbitrary figure that every banker is happy to maintain.

There doesn’t have to be a written agreement for collusion to exist.

Posted by KenG_CA | Report as abusive

> If US IPOs are used mainly to provide pricing

“There you go again” :-( Can you provide ANY argument, evidence, _other_ source than your firm personal conviction, to suggest that any participant in the IPO process has this even as as a secondary or tertiary goal, or even knows what it would mean? Bonus’s linked to price-discovery? Something in the prospectus or filed with teh SEC? Or maybe a revealed-preference argument showing that somehow IPOs are good at providing pricing (I would imagine this might involve an analysis of the size of the float, otherwise we could just list one share, right?)

I am happy to provide pricing, for the princely sum of $0.00. Here it is: $10/share. But of course, this is the _wrong_ pricing and an IPO (or just a properly run one??) provides the _correct_ (/better) pricing according to the following definitions and/or scholarly citation… [Felix, please help me out here?]

Posted by bxg5 | Report as abusive

Great post Felix…

I was in business school 10 years ago and my professors railed against the IPO process as one in which all parties except the investment banks get screwed.

High profile companies often issue at a massive discount to the market open. (like Linkdin) So the company get’s screwed. Investors have no idea how many shares they will be allocated… just that if they get lots of shares the deal is probably a flop and if the deal sells like hotcakes the small investors get shut out completely.

10 years later and the process still sucks.

why can’t the NYSE or the Nasdaq just run a Google style dutch auction for new listings using SEC audited financials?

Posted by y2kurtus | Report as abusive

All companies have a deep aversion to cutting prices or cutting quotes, but they typically cut prices if they have high fixed costs to recoup and their volume does not reach the level needed to recoup those costs. The classic example of this behavior is airline pricing. After AA tried to introduce a “simpler,” but in fact higher,” pricing structure in the 90′s, TWA and Continental undercut them very quickly in order to fill up their planes. As long as they made more than their per-seat variable costs, they would cut prices.

Theoretically, a company would cut their prices even when times are good. In practice, I just have not seen that happen. Companies do not generally do elaborate calculations of more volume for lower prices when they lower their price. They generally only do so when their volume starts to lag and they start losing money (mass-market retailers are an exception to this).

For IPO’s, you would think banks would have dramatically lowered their prices during the post-2000 IPO bust, but investment banks have some idiosyncrasies compared to companies that actually make stuff:

1. Their fixed costs are fairly low. They have office space and support staff, but for the IBD itself, these fixed costs are minimal compared to their revenue.

2. Most companies would consider compensation for their white-collar and skilled employees to be fairly fixed. Even today, many managers would rather cut prices than cut compensation costs. On the other hand, investment banks have more built-in variability in their compensation structure. Analyst bonuses may range from 10k in 2001 to perhaps 80+k in 1999 or 2006. There were stories in 2001 of IBD analysts working for less per hour than store managers at McDonald’s.

Banking managers also tend to have less reticence about layoffs during bad times. In this respect, their compensation costs are less fixed.

I would guess that in both respects, variable vs. fixed pay and willingness to lay off people in bad times, European banks’ compensation costs are more fixed. In bad times, they would rather lower prices to recoup compensation than lower compensation and keep the same costs.

I also wonder why the current IPO system has not been completely commoditized. My only guess is the role of gamemanship with IPO’s. In the regular underwriting system, Goldman Sachs sells to Fidelity who then probably gets a nice first-day gain, or “pop.” It would seem this first-day gain would hurt the capital raised by the end company, but in reality it helps publicize the stock for brokers to hawk it to small retail investors. With the publicity at their disposal, mutual funds can now capitalize on the pop through using an army of retail brokers.

To put it in more crude terms, the system is basically built to defraud the least possible rational investors, the retail shops. Eugene Fama may not approve, but that’s how the real world works.

The question then becomes why couldn’t mutual funds do the same thing with auctions instead of a 7% underwriting broker? The issue is that after the bank sells the IPO to its funds, the very next trade becomes the published price. The very next trade is whoever has the highest bid, or the least rational person, and therefore you get a big pop from the bank’s price to the “market’s” price, when the market is whoever the highest bidder is.

For auctions, that highest bidder is bidding alongside the funds. There isn’t the publicity of the funds buying the stock and then the “market” saying that it’s really much more valued. Without that publicity, the retail brokers do not have as good a story to tell. You would think they could tell the same story before an auction and the underlying company could get the value instead of the mutual funds, but it doesn’t work that way. Brokers just don’t have a good story for companies with no price discovery and that feeds down to a lesser price in the initial offering.

Posted by mwwaters | Report as abusive

Might be a good idea to add a caption explaining that that thick black line is *not* a best-fit curve or average line or something, but is the data itself.

Posted by gwern | Report as abusive

Let’s see now. The data for both the European and the US distributions appear to be homoskedatic in deal size — so that all the information to be had can be derived from the projections onto your Y (percentage) axis.

There can be little doubt that IPOs are more expensive in the US. But to say that the US has a cartel whereas Europe does not, you would have to sow that the US distribution is significantly narrower than the European distribution.

True, the range is clearly greater for the European than for the US distribution; but there are so many more points in the European distribution that this can easily be misleading. Eyeballing the plot, I think the US distribution has a raw std deviation of about 1% and the European distribution has a raw std deviation of about 1.5%. These should be readily computable. Applying a statistical test (probably some sort of Chi-Sq test), my guess is that you would not find the two distributions to differ significantly in variance, and that the hypothesis that the US has a cartel whereas Europe does not would therefore not be borne out.

It would, of course, remain to be explained why the US is a more expensive place to do an IPO than Europe is, but that’s a very different question, best answered by asking someone who id an IPO in the US, “Why did you do your IPO in the US, when it’s so much cheaper to do it in Europe.” My guess is that most of the answers would be, “Well, our VC has a long relationship with bank X and he felt very comfortable going with them.” Comfort is expensive, as anyone who has purchased an automobile is well aware.

Posted by samadamsthedog | Report as abusive

“True, the range is clearly greater for the European than for the US distribution; but there are so many more points in the European distribution that this can easily be misleading.”

You misunderstood that chart — the thing that appears to be a solid black line is in fact the 95% of US data points that are EXACTLY 7%.

Posted by starwed | Report as abusive

Fascinating article — and very damning.

It reminds me very much of the real estate sales market. Realtors® charge 6% regardless of costs or conditions and they do not compete on price.

Recently, there have been a few cracks in the armor of real estate sales — low-commission realtors, limited-service, etc. But it’s not easy to break the 6% norm. The trick is that buyers care so much more about the ultimate price, which varies wildly — by much more than 6% or 7% — that realtors and banks doing IPOs can persuade their customers that they have more to lose by going to any emerging low-cost rival than they have to gain. Sure, your commission may be cut in half, but your price will be off by 10% or more. So why do it? (And that’s assuming there even _are_ low-cost rivals in the IPO market, which the chart suggests are very, very rare.)

These banks really are vampires, sucking blood from the rest of the economy. It is so disgusting.

Posted by JR04 | Report as abusive

I think the key thing here is the difference headline fees and real fees. Having worked in this area (admittedly European debt not US equity markets) it was standard practice to quote a headline fee in the prospectus that higher than the one actually charged (ie part of the headline fee was rebated back to the client).

High headline fees are inevitably popular with the banks as they do not like to advertise discounts. It may be that this approach is simply more common in the US – not sure how/if the authors in the study adjusted for this.

Posted by cb22 | Report as abusive

Having worked at a bulge bracket IB and executed multiple equity transactions, this comes as no surprise. Firms will not compete on price. Full stop. Collusion isn’t limited to fees, banks coordinate the IPO calendar by sector to manage the supply of deals coming to the market. The underwriting process across all capital markets products represents an enormous profit pool for the banks, completely disproportionate to the actual risk taken in the execution. The bulge bracket firms collude to protect access to the IPO mkt and preserve fees. On the debt side, they collude to maintain pricing power. Fees are seeing as a payback for extending credit to issuers.

Posted by Herb1 | Report as abusive

Having worked at a bulge bracket IB and executed multiple equity transactions, this comes as no surprise. Firms will not compete on price. Full stop. Collusion isn’t limited to fees, banks coordinate the IPO calendar by sector to manage the supply of deals coming to the market. The underwriting process across all capital markets products represents an enormous profit pool for the banks, completely disproportionate to the actual risk taken in the execution. The bulge bracket firms collude to protect access to the IPO mkt and preserve fees. On the debt side, they collude to maintain pricing power. Fees are seeing as a payback for extending credit to issuers.

Posted by Herb1 | Report as abusive

Numerous smaller middle market banks are willing to bookrun deals at a lower underwriting spread. I’m going to go out on a limb and guess that Cowen isn’t going to be picked by Facebook to bookrun their IPO. The CEO’s and entrepreneurs of these Companies apparently see value in having a “brand” name bookrun the deals and are willing to pay a higher spread for the privilege.

Additionally, if you abolished the SEC and the registration process, and let Companies freely issue shares in a manner they saw fit, issuance costs would drop.

Posted by TinyOne | Report as abusive

Posted by KenG_CA:”There doesn’t have to be a written agreement for collusion to exist.”

Exactly. Collusion in the Madoff case, the mortgage securitization and MERS, and forclosuregate for 3 of thousands.

Posted by TinyOne: “Additionally, if you abolished the SEC and the registration process…”

That simplifying is how you get a MERS mess… so I doubt that would be the ticket to anything but more hell.

Thanks for asking hard questions Felix.

Posted by hsvkitty | Report as abusive

some options for companies seeking capital might be:
- perform a “reverse tender offer” where unlike a takeover artist seeking control of a target company by asking people to tender their shares for a set price, they could register securities and go directly to the public through an auction process. If they want to sell X share respresting x% of their capital, get prospective buyers to submit bids

- or acquire a publicly traded company and do a reverse merger. Practically fee-free compared to an IPO. But doesn’t generate a lot of publicity.

Probably the companies are really after publicity more than the investment capital… so they can say “advised by Goldman Sachs” etc. etc.

Posted by smokeygeo | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/