But then I read Herbalife’s press release and it turns out that there’s a good answer to that question, which is that you don’t have to borrow the stock and sell it short at all. You can just enter into a contract to sell the stock to Herbalife in five years, which comes to roughly the same thing — you’re short the stock through that contract, rather than in the traditional way. So you don’t have to worry about borrow costs, and Herbalife can have its convertible. Also it can buy back a lot of stock, which was really its goal.

The way it works is that Herbalife enters into a five-year prepaid forward contract with one or more of its underwriters — Bank of America Merrill Lynch, Credit Suisse, HSBC and Morgan Stanley — to buy back some of the shares underlying the convertible. (Five years because the convertible matures in August 2019, so really five and a half years.) Let’s say that’s 9 million of the 12 million shares underlying the convertible. 1

Herbalife gives the banks the money now — say it’s about $590 million 2 — but it gets its shares in five years. 3 Because the forward is prepaid, Herbalife gets to take the shares out of its share count for earnings-per-share purposes now, so for accounting it’s as though it had done the buyback right now. But in fact, the banks have the money and Herbalife doesn’t physically have its shares. If that sounds like an odd deal, it’s because it is. 4

The banks then enter into five-year swaps with the hedge funds buying the bonds. 5 So the bank in substance owes Herbalife 9 million shares and is owed 9 million shares by the hedge funds. (The bank is flat, because the bank is always flat, that is its nature.) The hedge funds are, collectively, long $1 billion of bonds (with 12 million shares underlying) and short 9 million shares. I sort of made up that 9 million shares, but it’s meant to represent the theoretical hedge ratio of the convertible: The hedge funds have models to tell them how many shares they need to short to hedge this convertible, and my guess is that the answer is around 9 million. 6

There’s also a capped call, about which the less said the better, and the more said the footnotier. 7 In brief, Herbalife is effectively raising the conversion price of the bonds by entering into an option with the banks, which the banks need to hedge by buying even more shares. To do this, they expect to purchase the common shares over a five trading day period immediately following the pricing of the Convertible Notes.

I have created a diagram-like object that sort of works if you don’t think too hard about it:

Can you guess what I used to draw this? I’ll give you a hint: Never use it to draw things.

So Herbalife is raising $1 billion and using the bulk of it on immediate stock buybacks, some of which comes from market sellers but most of which comes in synthetic form from the hedge funds buying the convertible. Pretty neat.

After today, the hedge funds sit around for a while, adjusting their hedge based on what the stock does. In five years, if the convertible is in the money (the stock is above $83ish), the hedge funds convert it, get 12 million shares, deliver 9 million of them to the banks to close out their swaps; then the banks deliver them to Herbalife to close out the forward, everyone is happy. 8 If the convertible is out of the money, the hedge funds get their $1 billion back, use it to buy 9 million shares (for much less than $1 billion), deliver the shares to the banks, who deliver them to Herbalife, and the hedge funds keep the rest of the money. Either way, it all works out fine.

Why do this? For one thing, Herbalife gets to do a giant gob of buyback all at once: They certainly have plenty of cash flow to fund a buyback effort, but by taking on debt that gives them the ability to accelerate the effort, as one analyst puts it. Of course accelerate is a loose term — they get the shares in five years! — but for economic and accounting purposes it’s true. They’re buying 9 million shares or whatever at today’s price, and retiring the shares for earnings purposes today.

For another thing, normally when companies offer convertible bonds, their stock goes down, because hedge funds are selling short to hedge the convertibles they’re expecting to buy. On this deal, though, no one is selling to hedge the convertible — Herbalife is, in effect, buying from anyone who wants to sell. In fact people are buying: The banks are buying some for the capped call, and depending on the exact structure, banks or hedge funds might be buying to hedge the forward too. No wonder that Herbalife started today up significantly (also: decent earnings ), though it’s given that away and then some as the day’s gone on.

Also, it makes the convertible deal more attractive: Instead of paying whatever Bill Ackman and the other fundamental shorts are paying to be short Herbalife, the hedge fund investors are getting short at, presumably, a cheap rate through swaps with the underwriting banks. In fact, depending on exactly how it’s set up, they might be long some shares too, which they can lend to make a little extra profit. 9

What conclusions can you draw about Herbalife, Bill Ackman, etc. from this? 10 I don’t know; the point of this is mostly aren’t convertible bond deals weird? The expanded buyback is a sign of confidence from Herbalife, obviously, and the four-bookrunner convertible deal is at least supposed to be a sign that BofA Merrlll, Credit Suisse, HSBC and Morgan Stanley share in the company’s enthusiasm, or at least its belief that it is not an illegal pyramid scheme that will go to zero in the near future.

On the other hand, though, among the possible stories of someone spends a billion dollars to buy Herbalife stock, this seems like the most benign for Ackman. Unlike a lot of theorized share purchase transactions — Carl Icahn doing a tender offer. for instance, or William Stiritz doing a leveraged buyout. or even just Herbalife buying back a whole lot of shares on the open market — this transaction doesn’t actually reduce the number of shares that are physically outstanding. 11 And while those sorts of transactions would make stock borrow more expensive and squeeze short sellers, this one won’t. This trade is designed to create stock borrow (for convertible arbitrageurs), and so might end up making it cheaper for short sellers bet against Herbalife.

Most intriguingly, with banks looking for synthetic short sellers to hedge their big short position to Herbalife, you could imagine them being tempted to give short sellers like Ackman some swap exposure as part of their hedge. Presumably Herbalife would take as much stock as it can get on the forward. The banks will prefer to sell the convertible to hedge funds, who will want swaps, than to outright investors who won’t. And could you blame them if they were a little tempted to source some extra shares by writing swaps to outright short sellers? After all, that would both increase the size of their trade and get their client — Herbalife — what it wants, in the form of a bigger buyback. And that would really make this deal perfect, if it ended up allowing Ackman to cheaply short some more Herbalife back to Herbalife itself.

1 The rumor is that the range is a coupon of 1.5 to 2 percent and a conversion premium of 25 to 30 percent above the stock’s closing price today. Let’s say that price is $63 and that the deal comes at the midpoint of the range for a 27.5 percent conversion premium. That’s a conversion price of $80.33, for about 12.4 million shares underlying a $1 billion convertible. I’m going to say 12 million because it’s easier.

I guess that the delta — the theoretical hedge ratio — of that convertible is about 75 percent, based on some rough Bloomberg CVNI math. Seventy-five percent of 12.4 million is about 9.3 million shares.

That’s just the most logical number of shares to buy back — the delta of the convertible. Smaller numbers are very possible, if, for instance, some non-hedge investors buy the convertible, or if Herbalife just doesn’t want to buy back that much. Larger numbers are also possible, though more complicated; see note 5 for why.

2 That’s just 9.3 million shares times the assumed closing price of $63 is $588.2 million. It’s a forward so there’s some accretion going on but there’s no particular reason to worry about that here.

3 Though the banks can deliver the shares early if they want to, for regulatory/ownership reasons, for instance, or if the bonds are way out of the money and they’re bored, or borrow gets really easy and the whole thing seems pointless. But Herbalife can’t force them.

4 The fun, of course, is bankruptcy. Lehman Brothers went bankrupt with a number of these sorts of transactions outstanding: Lehman had the companies’ money, but they didn’t have their stock. Oops! That did not go well. (Though one was with Intel. which had the good sense to take collateral, but most companies don’t take collateral for accounting reasons, so it tends to go more poorly.) Here Herbalife seems to be spreading the risk among four banks so should be pretty safe, but five years is a long time.

Since I have you here, by the way, there’s bankruptcy risk the other way. Lawyers worry about share repurchase transactions where the issuer goes bankrupt: If BofA Merrill has $145 million of Herbalife’s money (1/4 of this deal) and owes Herbalife 2.3 million shares, and Herbalife goes bankrupt and those shares become worth $0, will a court just say okay, hand over the worthless shares? Or will it find some way to conclude that Herbalife was insolvent at the time of repurchase and demand the $145 million back? The fact that these underwriters are doing the share repurchase is a vote of confidence in Herbalife’s financial viability.

Of course I guess so is underwriting the deal in the first place.

5 I’m stylizing all the facts here a bit, and there’s some diversity in practice, but what I’m describing should more or less work and is reasonably close to what is probably actually happening.

The thing in the text that is most likely to be wrong is that the forward repurchase could be for as much as for the full 12 million shares. In that case, either the banks would buy back the remaining 3 million shares themselves and give the hedge funds 9 million shares of swaps, or else the banks would do 12 million shares of swaps and the hedge funds would buy the 3 million extra shares that they don’t need to hedge. The advantage of this is that, if you need to increase your hedge ratio (because the stock goes up), you can do it by just selling some of the extra shares that you’re long. In the version in the text, if the stock goes up, you have to borrow and short stock, which is expensive. Of course if the stock is up maybe it’s because the shorts have been squeezed and the borrow is cheaper?

Anyway, they haven’t yet disclosed how many shares they are buying in this transaction so this is all somewhat speculative.

6 See note 1.

7 In the capped call, Herbalife buys an option from the banks that mirrors the option in the convertible — so, say, it’s a call option on 12 million shares at a strike price of around $83, same as the convertible. Then it sells to the banks a call option on the same number of shares, but at a higher strike price. So in effect Herbalife only issues shares if it goes above the higher strike price.

I have not yet seen any disclosure on the terms of the capped call, but if we just pretend that the upper strike is up 75 percent, then monkeying around in OV gets a price for the capped call of about $11.11 per share, or about $138 million, and a delta of about 18.6 percent, or about 2.2 million shares (around $146 million worth). Those numbers inform the diagram. But they’re much more fake than any other numbers in this post; they just struck me as quasi-reasonable.

8 Technically the bonds are net share settled — settled in cash and, if applicable, the company’s common shares — but it comes to the same place. Also the swaps are surely not settled by physical delivery but, again, comes to the same place.

9 Not that much — I hear that Herbalife borrow rates are around 1.25 percent.

10 Incidentally there’s a very different flavor of transaction — called a stock borrow facility — that has some essential features in common with this. Actually it’s the same trade except that, in addition to whatever else is going on, the issuer also sells the stock underlying the convertible, as well as buying it back on a forward basis. It’s a way to create borrow for hedge funds to hedge their convertibles, for companies that don’t want to actually buy back stock. It tends to go. less well than the buyback version. It is controversial. One reason it is controversial is that some people think it’s a way to facilitate, not convertible arbitrage, but people who want to short stocks and can’t get enough borrow. I think that’s mostly wrong but maybe not 100 percent wrong.

[Update: Intensely important terminological point here. There is a thing called a Happy Meal convertible offering. In my experience, that term is used by bankers to refer to the sort of deal that Herbalife is doing here: Issue a convertible, buy back some or all of the underlying stock in convenient swap format, and maybe do a capped call on top of it (not critical to the Happy Meal concept but typically goes along with it). But for some reason when the Wall Street Journal wrote about convertibles with stock borrow facilities, and how those deals went wrong for little companies like Energy Conversion Devices Inc. it referred to those (very different! though with some important features in common!) deals as Happy Meals. So there is much confusion in the land. But we can take back the term! This is a Happy Meal. Energy Conversion’s ill-fated borrow facility deal was not. You can tell because Herbalife is happy, and Energy Conversion is not.]