- Google has two classes of public shares, Class A (GOOGL), which have one vote each, and the nonvoting Class C (GOOG) shares. The shares are otherwise almost identical.
- GOOG has been cheaper than GOOGL, although the voting rights should not be worth much.
- GOOGL will be removed from the NASDAQ OMX, but not S&P, indices near the end of June, potentially reducing or even reversing the price differential.
- GOOG holders may receive an “adjustment payment” based on the price differences in one year.
- Easy long-short or pairs trading strategy: Long GOOG and short GOOGL until discrepancy narrows.
Google recently created a new class of publicly traded shares by issuing one share of its new Class C (NASDAQ:GOOG) nonvoting shares for each share of its existing Class A stock (NASDAQ:GOOGL). This recapitalization had two major effects. First, it split the stock and brought the nominal stock price down from low-earth orbit into the stratosphere. (There is still plenty of additional room for Google to do a 10:1 stock split, which I think they should do.) Second, Google can use its ability to issue more Class C shares as currency for takeovers and employee stock options without further diluting the voting power of the founders. (The Google founders Page and Brin own most of the Class B shares, which have ten votes each compared to one for the Class A shares, and thus have voting control of the company. According to page 22 of Google's 2014 proxy statement, Sergey Brin and Larry Page together control 56.1% of the total voting power. Eric Schmidt has 7.6%.)
Having two outstanding classes of essentially identical shares leads to a simple long-short pairs trading opportunity: Monitor the prices of GOOG and GOOGL, and when they get out of line, go long the cheap one and short the expensive one. Eventually the prices should converge, and then unwind the position.
Here is a graph of the relative share prices:
Source: Author's graph based on data from Yahoo Finance.
The real question is, converge to what? Should the share prices be identical? Or should there be a permanent gap between them? Are the nonvoting GOOG shares worth intrinsically less than the voting GOOGL shares, and, if so, by how much? Note that so far the median price differential between the two classes has been $7.48, or 1.43%. Is that the right differential?
Corporate voting rights generally aren't worth much in the United States. A recent Journal of Finance study by Kalay, Karakas, and Pant found that the market value of corporate voting rights around voting dates was in the range of 0.1% to 0.25% of the share price. They calculated this value by examining the value of the common shares with the theoretical value of a synthetic stock made with options.
Here is a graph from the Kalay et al. study:
Source: Kalay et al. study
A similar study in the Journal of Corporate Finance by Kind and Poltera found an average of 0.37% of the share price. An older study in the Journal of Financial Economics by Nenova that looks at the difference in share price between the voting shares and the lesser voting shares in dual class companies found a value differential of about 2%. However, results from dual class companies need to be interpreted with caution as often there are serious differences in liquidity between the different classes of shares.
Although the average results from academic studies can give us some insight, they don't provide a precise number to apply in the specific case of a triple-class company like Google. In the case of Google, even the Class A voting shares have inferior voting rights, and the founders have absolute control unless they have a feud. Only in the event of a splintering of the Page/Brin alliance, which seems very unlikely in the near future, could there be any situation in which the voting rights could possibly make any difference at all in a corporate election. One could argue that with their lock on the company, Page and Brin effectively have 100% of the voting power. This would imply that the Class A and Class C shares effectively have equal voting power (both equal to zero) and thus should sell for the same amount.
However, one can spin future scenarios in which Page and Brin or their heirs decide to sell shares for various reasons and their voting stake in the company falls. Note how Bill Gates has systematically reduced his stake in Microsoft as he devotes his efforts to his charitable endeavors. Thus, the voting rights in the Class A shares should someday be worth something, even though the present value of that something is not very much right now.
Here is one way to deal with the value of the voting rights: Let's suppose that the voting rights are worth 0.15% of the company, and that the company has a total market value (totaling all three classes of shares) of about $350 billion. 0.15% of $350 billion is approximately $525 million. Allocate 66.5% of that to the Class B shares based on the Class B shares fraction of the total votes and the remainder (about $176 million) to the Class A shares. Then divide by 281 million Class A shares and one gets about $0.63 per share for the Class A voting rights. If one were to adopt a 2% value for the voting rights, then the answer comes out to about $8.35.
I believe that the correct number is much closer to .15% rather than 2% as it will be many years before there is any possibility of the voting rights having any meaningful value. Thus any calculations should take the time value of money into consideration and discount the value of the voting rights sharply. Furthermore, Google is more likely to issue the nonvoting Class C shares for acquisitions or employee stock options, tilting the liquidity balance in favor of the Class C shares.
Actions by major index providers to drop one of the share classes from major indices could also affect the relative liquidity of the two share classes and thus the price differential between the two classes. Membership in major indices tends to make shares more liquid and thus more valuable to investors.
A previous version of this article reported incorrectly that S&P would drop the Class A shares from its indices, as it had previously announced. S&P had changed its mind and is now keeping both the A and C shares in the indices. I wish to thank the reader who pointed this out and to Seeking Alpha for taking prompt and appropriate corrective action. I also apologize for my omission in not catching this.
NASDAQ is apparently sticking with its earlier announced plan to drop the A shares from NASDAQ OMX indices as of June 23, although it remains to be seen whether it will follow S&P's lead and reverse its decision as well. Wilshire has apparently already dropped GOOGL from the Wilshire 5000 index as GOOGL does not appear in the Wilshire 5000 constituent list. Wilshire's normal policy is to include in its indices only the primary share for stocks with multiple share classes.
As at least one big index provider has shown a willingness to change an announced policy, watch for announcements as the index rebalancing dates in June get closer. If the Class A shares get dropped from one or more indices, there could be a shrinkage of the price gap between GOOG and GOOGL. A Journal of Finance article by Chen, Noronha, and Singal found that firms dropped from the S&P 500 do not drop in value, but that stocks added to the index increase in value. However, a more recent NBER paper by Chang, Hong, and Liskovich examines stocks switching between the Russell 1000 and Russell 2000 indices and finds price effects for both additions and deletions.
Liquidity and borrowing considerations
Any long/short strategy has to take into consideration the liquidity of the various share classes. As the new Class C shares were issued to all of the old Class A holders, both classes of stock are widely held by both retail and institutional investors, and thus easily available for borrowing. Trading volumes are currently about the same on both classes. I see no risk of a short squeeze affecting one class relative to the other.
It is possible that at some point down the road, Google might split the new Class C shares to bring the nominal price per share into a more normal trading range, while leaving the older voting Class A shares unchanged. This would tilt the liquidity even more in favor of the Class C shares and thus impact the price differential between the two share classes. However, such an action is probably a long way off.
Adjustment Payment Coming in a Year
One of the interesting sweeteners for the new Class C shares is that GOOG shareholders may receive an adjustment payment if the Class C shares trade too far below the Class A GOOGL shares. This arrangement was the result of the settlement class action suit that delayed the recapitalization deal. Here is how it works: Over the first year of trading of the new shares, the volume-weighted-average prices (VWAPs) for both GOOG and GOOGL will be calculated. If the VWAP over the year for the new nonvoting GOOG shares is less than the VWAP for the old GOOGL shares, then the shareholders owning the new GOOG shares on the one year anniversary of trading may receive an adjustment payment as follows:
If GOOGL VWAP =
Payment per GOOG share
Less than 1 %
1 - 2 %
20% of dollar Delta VWAP
2 - 3 %
40% of dollar Delta VWAP
3 - 4 %
60% of dollar Delta VWAP
4 - 5 %
80% of dollar Delta VWAP
More than 5%
Source: Author's calculations
So far it looks like the adjustment payment might be in the $1.50 range, but I would not be surprised if the differential shrinks to the point where the adjustment payment becomes zero.
Disclosure: I am long GOOG. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: I am also short GOOGL as part of the pairs trading strategy that is long GOOG and short GOOGL as described in this article. I may undo this trade at any time as described in the article if the differential in price between the two shares narrows significantly.