Special Situations: cheap voting rights
Set-up CMCSA/CMCSK; DISCA/DISCK; LMCA/LMCK
CMCSA/CMCSK: 1.005; DISCA/DISCK: 1.017; LMCA/LMCK: 1.005
August 5, 2014
We recommend setting up the above trades justified by mean reversion and the theoretical view that voting stocks should be worth a lot more than non-voting stocks, particularly when both classes pay the same dividends or corporate actions as is the case in the above situations. Moreover, in the three cases, the companies are controlled via a third class of super voting shares, the B class, which hardly trade or not trade at all. John Malone controls Liberty Media (LMCB) and Discovery Media (DISCB). The Brian Roberts controls Comcast (CMCSB). Liquidity, often a reason for price discrepancies between voting and non-voting shares (the liquid shares at a premium) cannot justify the lack of premium for the voting shares, except in the case of Liberty Media. However Liberty has only started to trade in its present form for less than 2 months (table below):
Daily volume average: voting (A) non-voting (K)
Comcast 4,000,000 600,000
Discovery 500,000 135,000
Liberty Media 220,000 450,000
The reason for the price dislocation may be found elsewhere. In the case of Comcast, it could be argued that the Comcast/Time Warner Cable deal is putting short pressure on CMCSA as this is the stock currency used in the exchange offer. The A shares started to underperform the K (or A special) ever since the transaction was announced last February. However, we note that previous rounds of underperformance (or outperformance) were not deal related. In any event, as the Comcast/Time Warner Cable deal completes, the short selling pressure on the A shares vs the K should abate and the premium on the A shares should re-widen. We think a position long A/short K should be built –up now because of the limited downside, the “ticking of the clock” on the deal. Moreover, the position long CMCSA/short CMCSK is a de-facto free insurance on a deal break (which we don’t anticipate; nonetheless the insurance is free). In the case of Discovery, the C shares (ie DISCK) built up a premium starting May 20 when the company announced a stock dividend in C shares to all shareholders, thus boosting the amount of shares outstanding in the C class, supposedly to increase liquidity in that class. The ex-date for this event is August 7, at which point we would expect holders of the A shares receiving C shares (DISCK) to sell them for the voting stocks given the small premium. For Liberty Media (post Liberty Broadband spin-off), the trading period is too short to form a view. We note that when the two shares started to trade following the spin-off of non-voting shares to voting shareholders, the voting shares were at a 6% premium, which then went to 10% before collapsing to near zero.
The Comcast and Discovery share classes are extremely mean reverting (next pages). We recommend setting up share for shares.
We would also advise that regarding the Time Warner Cable/Comcast deal, the risk arb deal should be set-up shorting the K shares rather than the As. In doing so, one implicitly sets-up a long TWC/short CMCSA position (as per the deal structure) and a long CMCSA/short CMCSK position. The profitability of the risk arb trade would therefore be enhanced by the expected pay-off on the share class arb, which by the time the deal is over could be an additional 150-250bps