The weird thing is, it looks like this every day. Photographer: Jin Lee/Bloomberg
The weird thing is, it looks like this every day. Photographer: Jin Lee/Bloomberg

You can learn a bit about an initial public offering by looking at the first day's trading. Here is a chart of how Candy Crush Saga maker King Digital's stock traded today:

This shows how many shares of King Digital trade at various prices. You might focus in particular on three noteworthy prices:

  • $22.50 is where King's IPO priced: Everyone who bought shares from the company bought them at $22.50. There is a bar at $22.50, but it is not visible to the naked eye: Roughly no shares traded anywhere near $22.50.1
  • $20.50 is the really tall bar; most of that is the opening cross at 9:51 a.m., for 5.1 million shares. Of the 25.5 million shares that King sold in this offering at $22.50, roughly one in five were then immediately re-sold at $2 lower.2 So one-fifth of the investors in the IPO put in the work, reviewed the prospectus, decided to commit to the deal, got allocated shares and ... re-sold them hours later at a negative 8.9 percent one-day return. Oops!
  • $20.00 is the other suspiciously tall bar; some 4.2 million shares traded there, versus around 1.5 million in the $19.90 and $20.10 buckets. That's an "extra" 2+ million shares trading at $20.3

Possibly that's because people and algorithms like round numbers but that seems unlikely. More plausible is that a lot of the buying at $20 was done by King's underwriters -- led by J.P. Morgan, Credit Suisse and Bank of America Merrill Lynch -- to support the stock price, using their 3.33 million share "greenshoe" overallotment option.

The name is dumb and not important, but the way it works is:

  • King sold the underwriters 22.2 million shares at a price of $21.21.4
  • The underwriters sold 25.53 million shares to the public at a price of $22.50.
  • The underwriters were short 3.33 million shares.
  • King also wrote the underwriters a 30-day call option, with a strike price of $21.21, to buy those 3.33 million shares.
  • So the underwriters could either buy back the shares in the open market, or exercise the option and buy them from King at $21.21.

The purpose of the greenshoe, from the selling company's perspective, is to support the stock. The banks are short the stock with the intention of buying it back if the price drops, which would provide buying pressure to cushion the selling pressure. The banks are a source of demand for shares if there's too much supply; they have "ammunition" to buy more stock if it falls below the IPO price.5

The purpose of the greenshoe, from the investors' perspective, is to support the stock: They don't want the stock to crack either! They're putting their money into a new public company; they want to know that someone else is a buyer too, even if it's just the banks who sold them the stock in the first place. Basically, no U.S. IPO gets done without a greenshoe, because investors want it.

The purpose of the greenshoe, from the banks' perspective, is to support the stock: They want their deals to go well, and will be sad if their investors lose money and blame them. But, as a little side effect, it is also a potential source of profit. Here, it is reasonable enough to conclude from that chart that King's banks probably covered the bulk of their short position at $19.00-$20.00.6 If you just casually assume that they covered the whole thing at $20, then the math goes like:

So the banks made $8.3 million supporting King today.7 And it only, you know, kind of worked; King closed at $19.00.

With the benefit of, like, eight hours of hindsight, you'd have to say this wasn't the underwriters' finest work. They sort of by definition priced the deal too high: After marketing at a range of $21 to $24, and pricing at the $22.50 midpoint, the market clearing price for King Digital turned out to be $19.00, suggesting that the company and its underwriters got a little greedy.

(Also uninspiring is the $20.50 opening price. Usually you use the greenshoe to support the IPO price: The underwriters buy more shares if the market threatens to go below the $22.50 IPO price. Here, the market never threatened to go below the IPO price -- it just went right ahead and did it at the open.)

Arguably this is a win for King Digital -- it sold its shares at 15 percent above what they're worth!8 -- but not really. King Digital didn't need the money that it raised in this offering: It IPOed because the offering "creates a liquid market for our current and future employees and equity holders and will give us greater flexibility to act on strategic opportunities if they arise in the future." An IPO that makes money for investors is better for those purposes than one that doesn't.9

The lesson here is: IPOs are hard! You have a thing that is worth who knows how much, and you want to transition that into a thing priced continuously by fast-moving public markets, and that transition is just going to be jarring and awkward. And you'll do your best to round up demand and guess the market-clearing price and sell it there, and sometimes you'll be 15 percent too high, and sometimes you'll be 73 percent too low, and that is life. Markets are unpredictable. You can't blame banks for this; one-fifth of King's initial investors immediately bailed on the stock at a $2 loss, suggesting that they had no idea what the right price was either.

So it's no wonder that banks do what they can to mitigate that unpredictability. They cajole clients into supporting weak deals with hints that better deals are to come, for instance. And they use the greenshoe option to support deals that go poorly -- even if sometimes, when that's necessary, it's already too late.

1 If you actually look at the price table (TSM), 500 shares traded at $22.50, out of a total of 40.9 million shares.

2 Technically King only sold 22.2 million shares, but the greenshoe is allocated: Investors buy the 22.2 million shares that King sells plus the 3.3 million, for a total of 25.5 million. This gets clearer further on.

3 In fact, that chart lumps trades into 10-cent buckets. If you look more granularly at the TSM screen, you get 2.27 million shares at $20.00 exactly, versus 313,054 at $19.99 and 271,564 at $20.01. So basically an extra 2 million shares at exactly $20.00.

4 That's the $22.50 selling price minus a 5.75 percent gross spread. (Per Bloomberg CACS.)

5 Is this market manipulation? Sure, if you want. But it is explicitly blessed by the Securities and Exchange Commission and by decades of practice, and is clearly and prominently disclosed. I don't want to hear it. Also explicitly disclosed, naked shorts:

In connection with this offering, the underwriters may engage in stabilizing transactions, which involves making bids for, purchasing and selling our ordinary shares in the open market for the purpose of preventing or retarding a decline in the market price of our ordinary shares while this offering is in progress. These stabilizing transactions may include making short sales of the ordinary shares, which involves the sale by the underwriters of a greater number of shares of ordinary shares than they are required to purchase in this offering, and purchasing ordinary shares on the open market to cover positions created by short sales. Short sales may be “covered” shorts, which are short positions in an amount not greater than the underwriters’ over-allotment option referred to above, or may be “naked” shorts, which are short positions in excess of that amount. The underwriters may close out any covered short position either by exercising their over-allotment option, in whole or in part, or by purchasing shares in the open market. In making this determination, the underwriters will consider, among other things, the price of shares available for purchase in the open market compared to the price at which the underwriters may purchase shares through the over-allotment option. A naked short position is more likely to be created if the underwriters are concerned that there may be downward pressure on the price of the ordinary shares in the open market that could adversely affect investors who purchase in this offering. To the extent that the underwriters create a naked short position, they will purchase shares in the open market to cover the position.

Every so often someone will tell me how outrageous this is and I will nod politely and pretend to listen. It is a thing. There it is. It's in every deal.

I have no idea if there were naked shorts here, incidentally. Sometimes if you're worried about a deal you'll put on a naked short, but it's a little weird for a deal to price at the mids (King's marketing range was $21 to $24 and it priced right at $22.50) and have the banks be worried enough about it to put on a naked short. But who knows? Maybe the banks wanted to price at $21, King pushed them to $22.50, and the banks threw on a big naked short to try to protect the price. (And/or to profit when the deal cracked, of course.)

6 There's a cluster of buying around $19.60, though it's not as concentrated at $19.60 exactly. There's another big cluster at $19.00 exactly, but some of that is just the close.

7 If they exercised the greenshoe, they'd have made $4.3 million on it (buy at $21.21, sell at $22.50), so that's $4 million of "extra" profit.

8 Er, 11 percent; remember it only cleared $21.21 per share after fees.

9 Though, I mean, Facebook is just fine.

To contact the writer of this article: Matt Levine at mlevine51@bloomberg.net.

To contact the editor responsible for this article: Tobin Harshaw at tharshaw@bloomberg.net.