Click here for full PDF
Under Armour (NYSE: UA), the Baltimore-based athletic apparel and footwear company, is well-known to most as the aggressive underdog in the industry dominated by the behemoths like Nike and Adidas. Dazzled by its growth, investors quickly took notice of its stunning growth and returns, and assigned it high earnings multiples (its trailing TTM P/E is a whopping 72x as of this writing).
At the helm of the company is Kevin Plank, the founder and CEO of the company. Currently owning approximately 66.5% of the voting power of all outstanding shares, Plank has majority control to guide UA to his vision.
March of this year, UA announced that it will issue a stock dividend, effectively a two-for-one stock split, except instead of two identical Class A shares, investors will get one Class C share for each Class A share they own. Class C shares don’t have the voting rights that Class A shares enjoy. The “change will [allow Plank] the flexibility of selling these non-voting shares of Under Armour over time while maintaining [its] founder-led approach.”
Shareholders didn’t seem too happy with the deal, evident in the large spread on the returns of the shares (shown in Figure 1). Note the heavy selling of both classes following the dividend.
Figure 1. 6-Month Returns on UA Class A vs. Class C Shares
In this report, I argue that there is a long/short opportunity with these two shares. Specifically, that the spread between the two shares are excessively wide, and the spread should narrow over time. Furthermore, as voting power is essentially concentrated in Mr. Plank’s hands, the premium of a shareholder’s vote should not be as high as the market believes it should.
Breaking It Down
Under Armour has three classes of shares: Class A, B, and C. Below is a summary of the share structure :
- Class A shares are publicly traded, and have one vote per share.
- Class B shares are wholly owned by Plank and his Family entities, and have ten votes per share. Class B shares are also convertible to Class A shares, and once converted cannot be reissued. Class B shares are not publicly traded.
- Class C shares are publicly traded, and have no voting rights.
- All three shares have equal rights to dividends, liquidation proceeds, and considerations in special events
There are finer details surrounding this structure. Most importantly is that each Class C share gains one vote on the conversion of all Class B shares into Class A shares. And that event will most likely be triggered when Plank and his entities together own less than 15.0% of the total number of Class A and B shares outstanding, which immediately converts all Class B shares into Class A shares.
To sum up, once Plank and his entities’ collective interest in the company dips below 15.0%, all outstanding shares convert into Class A shares, regardless of their previous class.
As of June 30, 2016, Plank owns 15.9% of all Class A and B shares, and 65.3% of the combined voting power.
As of June 30, 2016 there were 183,388,910 shares of Class A Common Stock, 34,450,000 shares of Class B Convertible Common Stock, and 219,454,106 Class C Common Stock outstanding.
Value of a Vote
Given that Plank and his entities own a majority control of the company, is the public shareholder’s one vote per share valuable at all?
If not, Class A share should, for all intents and purposes, be equivalent to Class C shares. Given their price disparity, this implies Class A shares are overvalued and Class C shares are undervalued, relatively speaking.
If investors do find the one vote per share valuable, the question becomes “how valuable is it?” The authors of a Harvard Law School paper, The Market Value of Corporate Votes, estimate the market value of the right to vote using options to create a synthetic long position. Similarly to non-voting shares, synthetic positions have economic interest in the company but do not have voting rights. The authors find that the mean annualized value of a voting right to be 1.58% of the underlying stock price in the US.
Using a similar method, I derived the value of a synthetic stock using put and call options. Specifically, a synthetic long position in Under Armour’s class A shares can be created by going long a call option and short a put option at the same strike price, as well as a long position in a risk-free zero-coupon bond that expires with a value of the strike price at the expiry date of the options. This relationship is derived from the put-call parity formula, and can be summed up neatly with the equation below:
This relationship is modelled with European calls and puts, so for the American options I’m working with, the equation should really be:
While this doesn’t provide exact pricing, it nevertheless establishes a minimum lower bound.
Using October 21st closing prices of November 16 call options, put options, and the underlying stock price, and assuming an annual risk-free rate of 0.66%, I get the following results:
Underlying stock price: $37.94
Strike price of $37.5:
Strike price of $38.0:
The difference in values come from using American options, as the option of early exercise has unequal effects on values with different strike prices, as well as the bid-ask spreads reflected by my broker. Nevertheless, it’s a good approximation of the value of a synthetic long position in UA.
These values are roughly in-line to the average value of 1.58% that the Harvard study found (1.58% and 1.74% for strike prices of $37.5 and $38.0, respectively). The options market evidently does not place too high a premium on a shareholder’s vote in Under Armour’s shares.
If a synthetic long position in Under Armour’s Class A shares is given only a modest discount for not having the privilege to vote, and Class C shares are similar to a synthetic long in sharing economic interest but lacking a right to vote, then the Class C share’s discount (or Class A share’s premium) is excessive and the spread should converge over time.
An argument for its significant lower price is could be due to downward pressure from Kevin Plank’s selling. On one hand, this makes sense – if there’s more sellers than buyers in the market, the price of the shares should go down.
On the other hand, this pricing disparity should not last for very long. Shrewd investors can find a bargain in Class C shares, buying an equal interest in the company for cheaper than Class A shares. The loss in voting power doesn’t matter too – Plank has majority control, and in the scenario that he relinquishes this control, Class C shares will be converted into Class A shares.
Given this price disparity, I recommend going short UA (Class A) and long UA.C (Class C), with an equal amount of shares on each side. I believe the spread between the two stocks will converge over time to a reasonable level. The uncertainty lies in the amount of time it will take for this spread to narrow.
T=0, UA: $37.88, UA.C: $33.00
Short 100 shares of UA, long 100 shares of UA.C
Net cost: cr. $488
T=1, UA: $40.00, UA.C: $39.20 (Scenario 1)
Exit short 100 shares of UA, exit long 100 shares of UA.C
Net cost: db. $80
Net return: 13.19%
T=1, UA: $35.00, UA.C: $33.95 (Scenario 2)
Exit short 100 shares of UA, exit long 100 shares of UA.C
Net cost: db. $105
Net return: 10.48%
Otherwise, an individual who wants to invest in Under Armour but finds the P/E multiple of Class A shares too expensive at a P/E of 72x might find the P/E multiple of Class C shares at 63x more reasonable (although given the aggressive scale UA is building, P/S multiples might be another way to judge relative pricing). Figure 2 demonstrates the relative cheapness of Class C shares compared to Nike’s shares.
- As Plank and his entities is only permitted to sell a limited amount of shares, the abated selling pressure should allow for the price disparity to slowly converge over time.
- Plank is nearing the 15.0% threshold (currently 15.9%) that would convert all Class B shares into Class A shares. The conversion would be a significant catalyst in effectively converting Class C shares into Class A shares. This is unlikely to happen in the short-term, as Plank still has a sizable position in Class C shares.
- Continued selling pressure by Plank and his entities may keep or widen the price disparity for an extended period (Plank currently owns approximately 15.4% of all outstanding Class C shares).
- Conversion of all Class B shares may not happen on a timely basis.
Because UA trades at high multiples, disappointing Wall Street’s expectations will serve to exacerbate selling pressure on both stocks, and hinder the price discovery process in the short-term.
- Lack of a clear catalyst (e.g. repurchases) that would cause the spread to narrow.
- Given the lack of a clear and timely catalyst, spreads could take a long time to narrow and significantly diminish time-weighted returns.