• 17 Jun 2013 at 4:13 PM

Let’s Hack Google’s Share Split

How much would you pay for a share of Google Class C stock? Those are the zero-vote shares that will soon be distributed, on a one-for-one basis, to holders of Google’s low-vote Class A shares, assuming that the settlement Google announced today goes through. We previously discussed the split when it was announced last year: Google’s founders don’t want to ever lose voting control of the company, so they’re proposing that any new shares issued for acquisitions, etc., be non-voting C shares; shareholder lawsuits have held this up until now but with the settlement it should go forward.

The traditional answer is that a share without voting rights is worth less than a share with voting rights because, y’know, sometimes you want to vote, and so various studies find something like a 2 to 10% discount for non-voting shares. But with Google that’s a little silly since no one really votes anyway: the high-vote Class B shares, which are mostly owned by co-founders Larry Page and Sergey Brin, give them about 56% of the vote, so whatever you do with your piddly Class A shares doesn’t matter. So the As and the Cs are basically the same except the As come with the hassle of having to mail back your pointless proxy card.1

So if you could get a C cheaper than an A, that seems like an arbitrage and you should buy it because the prices should eventually converge. But markets can remain irrational etc. etc. etc., so absent any obvious catalyst for convergence why would you do that? So Google, and some clever plaintiffs’ lawyers, provided a catalyst:

After the Class C shares have traded publicly for twelve months, the Board of Directors will compare the volume-weighted average trading price of the Class C shares for the initial twelve-month period (“the C share price”) with the volume-weighted average trading price of the Class A shares during that twelve-month period (“the A share price”). If the C share price is within one percent of the A share price, then no adjustment occur. If the C share price is one percent or more lower than the A share price, then the Board shall issue consideration to the then-current Class C shareholders. The consideration may be in cash, Class A shares, Class C shares, or a combination thereof, in the Board’s discretion. The amount of payment shall be made pursuant to the following formula:

  • If the C share price is equal to or more than one percent, but less than two percent, below the A share price, twenty percent of the difference;
  • If the C share price is equal to or more than two percent, but less than three percent, below the A share price, forty percent of the difference;
  • If the C share price is equal to or more than three percent, but less than four percent, below the A share price, sixty percent of the difference;
  • If the C share price is equal to or more than four percent, but less than five percent, below the A share price, eighty percent of the difference.
  • If the C share price is equal to or more than five percent below the A share price, one-hundred percent of the difference, up to five percent.

The idea here is that if the Cs fall to 1% below the As, then you can sell your A for 100x, buy a C for 99x, and also get a free 0.2x from Google, so you would. And the more out of whack they are, the more valuable the top-up is, which should also push them back together. Which, in turn, should make it unnecessary for Google to pay out anything on the top-up. The point of this mechanism is to arb itself out of existence.

A fun exercise is: can you break this?2 The drafting is sloppy enough3 that you probably can; I’ll give you one free idea though it’s not all that practical. The comparison, for computing the payment, is between “the volume-weighted average trading price of the Class C shares for the initial twelve-month period” versus the VWAP for the As over the same period.4 That is a mistake! Here’s why:

  • Buy a bunch of C shares & short the same number of A shares, so you’re neutral to the actual stock price.
  • Wait like 9 months and see what happens.
  • If this happens:

  • Then you might consider adjusting your A short obsessively – like, buy a lot of shares each morning and sell them back each afternoon – while sitting tight on your Cs.
  • On the other hand if this happens:

  • Then you might consider adjusting your C long in the same obsessive manner, while sitting tight on your As.

The goal is to manipulate volume rather than price: have relatively more A shares trade at high prices and C shares trade at lower prices, in order to create a “discount” for purposes of Google’s calculation. Then Google pays you for that discount on your C shares.

This is just for fun and probably doesn’t really work, of course, just because Google is so big. You’d have some slippage – you’d pay bid/ask in your obsessive trading – and you’d need to own, like, a lot of Google shares to make that worth your while. What guards this mechanism against manipulation is just the sheer size of Google.5 Though I guess that could be said about the FX market, etc. Anyway, leave your better ideas in the comments.

I suppose this is how the plaintiffs’ lawyers who sued Google earn their fees. We talked last week about how it’s genuinely hard for a controlled company to successfully navigate a merger that its controlling shareholder wants: on the one hand, the board has an obligation not to sell out the minority shareholders; on the other hand, the controlling shareholder is, y’know, controlling. Much the same dynamic applies here: since Google’s founders control more than 50% of the votes, it’s a little hard to stop them from doing what they want to do, which in this case is more or less “get even more of the votes.” The lengthy preliminary proxy for Google’s stock split bears this out: basically a lot of meetings were had! And then the board did what Page & Brin asked for, even though it’s sort of obviously shareholder-unfriendly in the abstract. In the less abstract, it doesn’t matter, it’s just about voting rights, and nobody really had those anyway.

But shareholder rights live in the abstract most of the time anyway, so there was a lawsuit, as there tends to be, and the judge made some annoyed noises, and so Google settled. But for what? They’re not going to not do it. The settlement has some governance-y protections, in which independent directors need to sign off on various nasty things that the founders could do in the future; you might ponder the fact that independent directors cheerfully signed off on this nasty thing too.6 Also it has attorneys’ fees, of course.

Also it has this mechanism for ensuring that the Cs and As trade at around the same level. Which, first of all: why? Each A shareholder now will end up with one A share and one C share when the split happens, so why should they care if they trade at different prices? This seems to be more about showing that the settlement did something than about actually preserving shareholder value. And, second: it’s a little broken.

Google Settles Non-Voting Stock Shareholder Suit [WSJ]
Memorandum of Understanding [EDGAR]

1. Oh I kid, why would you mail back your proxy card. Also there shouldn’t be a liquidity difference: day one there’ll be the same number of As and Cs. Eventually there’ll probably be a few more Cs since they’ll be used as an acquisition currency but it’s unlikely to get way out of whack just because Google is so big.

2. The obviously shady workaround of “buy As at 100x, sell Cs at 95x, and buy back separately the rights to receive the dividend after a year” doesn’t seem all that great to me but maybe it works? I find it less fun than the approach in the text but you could try it too. Obviously one question is who would sell you the dividend rights.

3. For one thing “equal to or more than one percent, but less than two percent, below the A share price” is hideously ambiguous. For another, “then no adjustment occur” is missing a word or something.

4. Pros would use the arithmetic average of the daily VWAPs over those 12 months, for reasons that will become apparent. Also for ease of calculation, generally.

5. This gives you some sense of the absurdity of the numbers:

That is, if the stock goes up linearly by 20% over that year, you’d need to like quadruple the daily volume of A shares at the end, while cutting the volume of Cs in half, to get a 3.6% fake discount (assuming no actual discount).

6. To be fair it also has a provision that “Google and its Board of Directors will not object to any judicial review being evaluated pursuant to the entire-fairness standard under Delaware law” for one of the potential nasty things (any board waiver of the provision where the founders need to sell non-voting and voting shares pro rata), which is rather clever; there was some debate in the current case as to whether the (harsh) “entire fairness” or (worthless) “business judgment rule” standard would apply.

2 comments (hidden for your protection)
Show all comments ↓

Comments (2)

  1. Posted by .Bo | June 17, 2013 at 4:25 PM

    3 articles before 5 pm? Multiple charts and graphs? Endless footnotes?

    I take it Breaking Media is no longer testing employees for PED's?

    -Willing to turn a blind eye

  2. Posted by spiderbro | June 18, 2013 at 12:28 PM

    My idea is short Google A and buy NSA for a delta neutral position