The Facebook IPO: An orgy of game theory in which everyone is screwed

So, I did my best to ignore the Facebook IPO during the frenzy which led up to the actual issuance of the shares. Sure I love Facebook, I'm on it all day. You're probably reading this through a link I posted to it. The start of my career in earnest was marked by Netscape, the first mind-blowing IPO of the internet boom. I went on to make my first career trading equity volatility in the telecom and media sectors during the first internet mania and collapse. I went on in my career to build infrastructure that supported the largest IPO in the history of the Middle East so I know what the IPO process looks like from the inside. Still, I have ignored the Facebook IPO and this is because, now that I know in excruciating detail how the entire process works, it disgusts me.

So now the inevitable has happened. The IPO process smashed into the hopes of the investors like a freight train and clipped them for 18% of their investment in the first week and 30% in the first two. It's hard to exaggerate just how big a fiasco this is. Over $4 billion of investor funds are up in flames and since this was meant to be “the peoples stock” with a massive 25% allocation to retail, mom and pop investors have been smoked to the tune tune of over $1 billion. Ouch, that will leave a mark. Indeed there have been a slew of articles talking about how the Facebook IPO has cooled the IPO market for other companies in the pipeline.

Now the recriminations begin. The journalistic community and the tort bar have both been beating the bushes to find the villains behind this who can be excoriated to drive page views or vilified and then shaken down for cash. This is entirely natural but I think it does a disservice to the public because it seeks to explain the outcome in terms of personalities, which granted are very interesting, but it allows the narrative to ignore the complexities of the IPO process itself. I think this is a disservice because in this case what has happened is not that some villains seized control of the process, indeed while some dubious decisions were made I'm quite certain that no laws were broken. No no, the problem here is not that some villains manipulated the process, the villain IS the process.

The important thing to keep in mind that this story is not about what Facebook is actually worth in terms of its balance sheet or income statement. No human being could possibly tell you that. Indeed, everyone with access to the US equity markets voting with their dollars for the past two weeks can't come to an agreement within $4 billion dollars on any given day. There have been stories about how some analysts lowered their estimates but didn't tell clients because of the quiet period and the implication is that some people got advanced notice and trimmed their bids. Let me assure you that anyone with privileged access to Wall Street research could remember the '99-2000 era and knew full well that the pricing and trading of Facebook was going to have nothing whatsoever to do with what the Wall Street analysts though 2013 revenues were going to be. The real value of Facebook has nothing to do with its revenues per user, or its compound annual growth rate, or any other measurable statistic. Everyone in the game recognizes that Facebook is a one off, a business that no one fully understands other than to know that it is huge with a ton of potential. Facebook is a dream of a company and the price of a dream is easy to calculate: it is worth what people are willing to pay for it. No more. No less.

Now I realize, dear reader, that this may seem manifestly unhelpful but behind what seems a free market tautology are the more disturbing questions about the Facebook IPO and therein the black magic lies: Who came up with the the IPO price and how did they do it?

The process by which an IPO is priced is called a “book-building.” Bookbuilding is a slow motion, mediated auction with several stages. The first is a filing with the SEC announcing the intention to list, and giving the fundamental details of the company and the offering as prescribed by law. Then there is the road-show where the bankers and the management travel around the country meeting with potential investors, answering their questions and feeling them out for their interest in the deal. It's kind of like dating where the company and its potential investors meet each other, tell jokes, buy each other dinner, and decide whether to take their relationship to the next level. Then begins the formal book building. The bankers have made a back of the envelope calculation of the company and come up with a range. In the case of Facebook that range was $29-$34. Then they invite people to submit orders to them.

There are several answers to the question “who.” First there are the potential investors who at the highest level can be broken down into retail and institutional investors. Usually retail investors are not invited to submit bids, their decision is generally a binary one: are they in or out at the price whatever it may be. The people who determine the price are the institutional investors. These in turn can be broken down into “hedge funds” and “real money” who are mutual funds and the big pension funds and insurance company funds. These institutional investors decide how much they want to participate in the IPO and then submit bids with a maximum price and for a specific amount of shares. These bids are received by second half the the “who:” the Equity Capital Markets (ECM) teams at the banks in the IPO syndicate who then manage the how: the “book build.”

As I have said a book-build is a mediated auction. By “mediated” I mean that the ECM teams, the “auctioneers” if you will, have sole discretion in the outcome of the auction both in terms of what the final price is and in terms of which investors get how many shares. An unmediated auction is what you might see at Sothebys recently when “The Scream” was auctioned off. The bidders turn up at a set date and time, and then they bid against one another and the highest bidder wins. In a book build the ECM teams collect all the bids from all the market participants and then they alone decide at their own discretion what the final price will be and which of the bidders will get how many shares, indeed this is what the company that hired them to conduct the IPO is paying them for. This process is highly opaque but I will try to illuminate it for you.

In business, as in life, always remember that people respond to incentives. If you can discern the incentive system in which someone perceives himself to operate you may as well have read his mind. So what are the ECM guys trying to achieve? They have several objectives. First of all they want the IPO to be a success and a successful IPO is one in which both the company and the investors are happy. This is important because if they shortchange the company they might have a hard time attracting future IPOs and if the investors aren't happy they might have a hard time selling future IPOs. That is, the ECM guys want to stay in business.

Beneath that they have some conflicting incentives with regard to the investors they select. In general they want the stock to pop a little on the first day to make the investors happy but not so much that the company feels as though it has left a lot of money on the table. To achieve this the ECM guys try to manage which investors get how many shares. They try to give large allocations to “real money” investors who will be in for the long term and small or zero allocations to “flippers” that is hedge funds who are mostly investing in order to ride the one day pop and then flip the shares. Within this there is also a certain quid pro quo wherein large clients of the syndicate member who pay a lot of fees and borrow a lot of funds and therefore pay a lot of interest will receive favorable treatment come allocation time relative to smaller clients and new clients had better be convincing that they intend to expand the relationship with the bank into other areas. So in sum, at the strategic level the ECM team needs to balance the price between the company and the investors in order to stay in the IPO game and on the allocation side they want to have the right mix of investors to support the price and they want to reward their high fee paying clients.

How about on the investor side? Basically there are two kinds of investors the “real money” guys like pension funds and insurance companies who are longer term in their outlook and who generally rely on the constant contributions from employees or policyholders to generate the funds they manage. Then there are the, mutual funds and hedge funds that have to compete with one another for investor assets to manage. “Real money” accounts are generally more patient investors and are mandated to match assets to future liabilities. Thus they tend to be more risk averse so they are more sensitive to price in IPOs. If they think the price is rich, they'll hold back and if they think the price is fair they'll participate to a greater extent. They have a lot of influence over most IPOs because the ECM guys think they're unlikely to flip the shares but because they are so sensitive to price they are less influential in a deal like Facebook. 

The incentive structure for the funds that have to compete for investor assets are much more complex. First of all remember how the managers are compensated. Generally mutual fund managers are paid a percentage of assets under management and hedge funds are paid a percentage of assets under management plus a share of the profits of the fund. This means that for both of hedge and mutual funds a primary driver for compensation is to have as many assets under management as possible and, for the hedge fund managers, if they can hit home runs while they have a lot of assets under management they can become dynastically wealthy. Next remember how they compete with one another for those assets, through a combination of risk and return, they want the highest return for the lowest risk possible and they need to produce better returns at lower risk than their competitors.

Now think about what this means for the incentive system in which the fund managers operate. IPOs, properly priced, are almost a sure thing investment, so everyone has an incentive to be involved. What's more some of the checks on the “real money” funds don't operate on the “fast money” funds. Remember that they are judged relative to their peers, so if everyone gets into the IPO and it performs poorly then the fund managers are no worse off, but if their competitors get in and they don't and the IPO turns out to be a massive win they are behind from both a performance perspective and a volatility perspective. So if it looks like there might be a pretty good IPO coming down the pipe a lot of fund managers will feel as if they must participate and therefore they are likely to be pretty aggressive in trying to get into them and to get as large an allocation as possible. It goes without saying that they are likely to be much more aggressive than the "real money" accounts in such a case. 

So this brings us to the Facebook IPO which has some unique features of its own. First of all is the nature of the company. It is a unique and possibly the most widely known company on the face of the Earth at the time of its listing. Given its spectacular growth, its short operating history, and the fact that is in a business which practically only existed as long as it has mean that arriving at a valuation based on comparables or history is virtually impossible. This means there are not a lot of solid benchmarks for analysts and investors to anchor their valuations to and that makes the mediated auction even more influential in determining price. Also it relaxes the constriction on the ECM team to get the price right. Since the IPO is so much different from all other IPOs an error one way or the other won't necessarily harm the future IPO business of the lead underwriter as other firms might consider this particular IPO to be a one-off.

If the constraints on the ECM team are relaxed a little on the valuation side of things they become massively more intense on the allocation side. For something like this anyone who is benchmarked against other technology funds ABSOLUTELY MUST participate in the IPO. If it turns into a huge home run and they are not in, their investors will wonder why. Did they just not have enough influence with the syndicate to get an allocation? Were they not smart enough to know that it would be a huge hit? How could that be? It's Facebook for crying out loud! On the other hand, since everyone has to be in, if the deal turns out to be a dud then all their competitors are smoked as well so they're not worse off on a relative basis. The thing is they have to get into the deal and they have to get a meaningful allocation. And this is where the game theory orgy takes over.

So let's say you want to get a large allocation in a mediated auction. How do you do that? Well, you can try to make sure you are the highest bid. So when Morgan Stanley announces the range is $29 to $34 you might want to be toward the top of that. Of course this is Facebook so you can probably guess that there will be plenty of people willing to pay $34 just to make sure that they're in the game, so you might submit a bid above the range. Actually, the syndicate will go you one better. They'll allow you to submit something called a “no limit” bid that means you will match the highest bid there is no matter what it is. Some nutcase is bidding $70 a share? You're with him. Of course you're assuming that whoever is the highest bid is less crazy than you, not always a safe assumption in this game and in any case you think the ECM guys won't price it too high lest it fall, also not a safe assumption. In essence,  it is so important to you to get an allocation of shares that you're willing to abdicate your influence over the price to marginally increase the chances that you get an allocation. Given how important it is for the competitive position of the various funds I'm sure a great many did precisely this. But it's not enough to get an allocation, you need to get a meaningful allocation. Lets' say for example that you submit a no limit bid for 100,000 shares and the IPO is oversubscribed many times, you might come away with 7,500 shares. That's nothing! You may as well not have been in the game. So if you want 100,000 shares, you might submit an order to 750,000 shares. Then, when the bankers go through their spreadsheets and decide who gets what now you might just get a decent sized allocation. And you know you won't get the full allocation because you know that the deal is sure as hell going to be massively oversubscribed.

Well so this is what happened. There the ECM guys are sitting in their conference room receiving bids all day long for weeks in front of the IPO. They are literally deluged with bids, a lot of them through the original price range. So many in fact that they raise the price range to $34-$38. Then they are deluged with “no limit bids.” So many that they can fill the whole offer at $38. Then they're also swamped with the sheer size of the share requests such that the IPO is “many many” times oversubscribed. And that's not all, not only are you overwhelmed with by the high price and the massive size you are absolutely swamped by requests from every client service person in the firm to look after their “very important clients.” It's hard to exaggerate the pressure the ECM team is under in the final days of the book-build to make everyone happy. And that's what the guys at Morgan Stanley tried to do.

They raised the target range from up to $34-$38 to make the company happy and in the end they priced it a the top of that range at $38. They almost certainly had so many no limit bids that they could have priced it even higher, I'm sure they thought they were leaving something on the table for the investors. They probably thought it would go to $45 or something on the first day just when they saw the demand. Indeed, even after pushing up the range they were still so overwhelmed with demand that they grew the offering size by 25% this enabled them to give their clients larger allocations and make all the whiney salespeople whine a little more quietly. I'm sure that it seemed all was well.

Alas, it was not so. Now imagine the position of our hypothetical hedge fund manager who, when the range was $29-34 submits a “no limit” bid for 750,000 shares thinking he'll get maybe 100,000. Well, now that they've grown the offer and, if he's done a lot of whining, he's probably wound up with 250,000 shares at $38. Hey now! That might be a little too many shares at far too high a price, but let's see how it trades on the open. Well, it trades up a little but there are these problems at NASDAQ and it starts trading off, what is he going to do? He's going to sell! And that is what has been happening, the crazy people who wanted as much as they could get at any price two weeks ago are getting shaken out and in their stampede for the exits they are crushing the living daylights out of the stock.

And that in a very long nutshell is what happened. There is no single villain or even group of villains. The nature of the incentive system in which fund managers operate compels them to bid with extreme aggression for access to an IPO of this kind. The very aggression with which they bid relaxes the incentives to constrain the valuation that operate on the ECM team and pushes them in the direction of allocating more shares to the institutional investors than they actually want which triggers a massive wave of selling when the post IPO demand fails to match up with the pre-IPO demand. Of course the Mom and Pop retail investors who just wanted to own a piece of Facebook have no idea how this all happens and most of them will probably hold onto their shares in the hope that things will turn up someday.

Meanwhile, as of todays close they've lost $1,450,000,000. Sorry guys, welcome to the big time.