Berkshire Ruminations

Friday, June 22, 2007

The Chipotle Mispricing Persists

The stocks of Chipotle Mexican Grill are both down today on an analyst downgrade, which is fine. The company’s price has certainly been bid up. What continues to amaze me, though, is the lack of attention to the very obvious mispricing that exists between the company’s class A and class B shares. For one thing, oftentimes the B shares aren’t even mentioned, or reports will just make reference to “shares of Chipotle,” as if they are all the same. Something very interesting has been going on here but no one seems to care.

When Chipotle was spun off of McDonalds last year in an IPO, it had only one class of publicly traded stock. McDonalds retained a majority interest in Chipotle, though, calling the shares it owned B shares and assigning ten times the voting power to them. Then, when McDonald’s sold its remaining interest in October (perhaps because it saw CMG as overvalued?) the B shares came to market.

The shares have the same economic value. That is, the B shares are identical to the A shares except for the fact that they have more voting rights. So, if anything, the B shares should trade at a small premium to the A shares. But they don’t. They have persistently traded at a 7% discount to the A shares. This is an example of investor irrationality. I read about it in various places after the phenomenon began, but it took a call to the company itself before I believed it.

Anomalies like this drive academicians nuts. How can this happen? Why aren’t arbitrageurs eliminating this discrepancy? The literature on behavioral finance contends that such anomalies can persist when there are “limits to arbitrage.” This explains why shares of 3Com traded at a discount to shares of Palm, which it owned, my favorite Wall Street anomaly of all time. Know-something investors couldn’t get their hands on shares of the overvalued stock to short, so it kept getting pushed up.

But I don’t think this can explain the persistent mispricing in this case. Both classes of stock trade several hundred thousand shares per day, and I personally have been able to buy both classes – the A’s before the B’s were around, then was able to easily sell the A’s and buy the B’s. So there is no liquidity concern here, and even if there were, I would expect the A shares to be discounted as there are slightly fewer of them outstanding (and thus they are less liquid).

So can this discrepancy be arbitraged? The short answer is yes, but it has not been profitable yet. Since the phenomenon continues to persist, those who shorted the A shares and bought the Bs have not yet realized any gain. And if it takes long enough to correct the mispricing, they may never realize a gain. I found it difficult to get historical price data for CMG.B because, again, everyone is focused on the A shares. It isn’t available on Yahoo, Google or even Ameritrade’s resources. I finally found a source for both classes on the Chipotle IR website, although I had to search for each day’s price independently. I then, painstakingly, charted the profit from an arbitrage of them by manually entering them into excel. And here it is, in all its glory:

I assume that I initiated the arbitrage trade on January 10, since this was around the time that I personally learned of the mispricing, by buying Class B shares at $52.40 and shorting Class A shares at $56.21 . The cost basis is the cost of the shares I bought, $52.40, so to date the arbitrage would have produced a gain of $2.00/$52.40, or 3.8%. If we assume the cost of shorting is 9% APR, or 4.5% over the six months we were short, then the interest on the short would completely wipe out the gain on the long. And this ignores the time value of money, the return on the market and, perhaps more importantly, the fact that a long position in either class of stock would have produced a great gain.

The analyst report that came out today also makes reference to the large short interest in CMG in the explanation of its downgrade. Couldn’t this simply be because of investors trying to arbitrage the difference?

PS - What do you suppose SocialPicks will say about this one?

FD: I have no position in any stock mentioned in this post.


  • I would suspect the short interest is due to the high p/e multiple, not this arbitrage. The problem is that you can't quantify what the spread on the two classes "should" be. This makes it hard to, one, know when to cover the short, and two, identify a catalyst that will narrow the spread. Comcast's spread has narrowed from its peak, but News Corp has a similar spread to CMG.

    By Blogger Chad Brand, at 22 June, 2007 13:41  

  • Chad,

    Whats your point? Obviously the spread between A and B shouldn't be negative, since Bs are clearly better than As. So although I can't quantify the magnitude I certainly can quantify the direction.

    Moreover, this example is completely different from both Comcast and News Corp. The Comcast spread is next to nothing and any profit would be wiped out from the commission alone - that is, its spread got corrected whereas the CMG spread has persisted from the get go. The News Corp spread, while larger, is understandable since the NWS shares are supposed to be worth more than NWS.A because of their superior voting power. That makes sense. CMG doesnt make sense.

    Something this obvious shouldnt need a catalyst, although apparently it does.

    By Blogger Andy Kern, at 22 June, 2007 14:04  

  • Mueller Water A and B shares is a similar situation that Geoff Gannon has written about. In that situation B shares have 8 votes and should trade at a slight premium.

    By Anonymous Anonymous, at 22 June, 2007 15:03  

  • How about a value player that can realistically give you 2000% return in 4 years, based on reasonable P/E of only 10, and the fact that palladium metal price doubled in past 2 years, and should double again in another two years.

    Do your own DD.

    By Blogger JJ2000426, at 24 June, 2007 20:43  

  • I agree with you that this drives academics nuts. As a fund manager, however, I use the A shares because they inculde the ability to hedge (b-shares don't have options or didn't when i began trading). This may not account for such a drastic difference but it at least is important to some investors.

    By Blogger Zach, at 25 June, 2007 15:33  

  • This comment has been removed by the author.

    By Blogger wimpyVO2max, at 26 June, 2007 09:17  

  • This comment has been removed by the author.

    By Blogger wimpyVO2max, at 26 June, 2007 09:22  

  • (posted this at seekingalpha)
    (pardon my deletions, cant get links right)

    Hey, if I was Dave Letterman I would call these "stupid chart tricks":

    A and B charts side-by-side:
    Chart 1

    Chart of daily ratio of A to B:
    Chart 2

    Chart 2 shows that the "arbitrage" is slowly beginning to close? currently at about 8%.

    By Blogger wimpyVO2max, at 26 June, 2007 09:29  

  • Andy,

    You are saying the B shares should trade above the A shares because of the voting rights? Isn't there a spread because of the B shares' connection to the parent company, whereas the A shares are the public shares? The market is assuming MCD will eventually sell off their stake.

    If you argue that there should be no spread (or it should be the other way) then sure, the arbitrage play is a great move. My point is that it is hard to pinpoint a catalyst for the spread to disappear as long as the dual class structure is in place. Surely the market would not be making a spread of this size if there was no reason for it.

    By Blogger Chad Brand, at 29 June, 2007 08:07  

  • Well, disregard that last comment. It looks like MCD no longer owns any CMG. The options point made before is a good one, but likely would not account for such a wide differential. Obviously there are some other reasons which aren't entirely clear.

    By Blogger Chad Brand, at 29 June, 2007 08:22  

  • I did short A class and buy B class in April, and I'm getting impatient. I think a big part of it is most investors don't know the difference, don't do dd, and assume that "B" is worse than "A". I wonder if the B shareholders can use their voting rights to enact a voluntary share recharacterization. I don't know if I can wait around, though, as the interest is eating up my expected gain. (Fortunately I did buy some extra B shares, so I am net long).

    By Blogger David, at 05 July, 2007 22:26  

Post a Comment

<< Home