Sirius XM (NASDAQ:SIRI) has had a remarkable run over the past year, increasing 121% versus a 30% appreciation in the S&P 500. The shares now sport a market cap of $11 billion and trades at an EV/EBITDA multiple of 17x 2011 Street estimates, near three times the multiple ranges of other subscriber based media companies.
For example, Dish Network (NASDAQ:DISH) trades at 6.0x 2011 EBITDA, DirecTV (NASDAQ:DTV) at 4.0x, Time Warner Cable at 6.4x, Cablevision at 6.5x, and Comcast at 5.8x (not factoring the recent NBCU acquisition).
As for the large cap media companies, who themselves have exposure to subscribers, CBS (NYSE:CBS) trades at 7.5x, Disney (NYSE:DIS) at 9.5x, Time Warner (NYSE:TWC) even with the recent run following the spectacular earnings report at 7.7x, Newcorp, a great company but with a depressed 5.0x multiple, and the pure play cable networks Discovery (NASDAQ:DISCK) and Scripps (NYSE:SNI) at around 10-11x.
Netflix (NASDAQ:NFLX) is the only other subscriber based company trading higher at 25x 2011 EBITDA. Both companies are at 20 million subs today but Netflix’s sub base is likely to triple over the next five years with international expansion, while consensus has Sirius XM adding about 8million subs over the same period.
Sirius XM is a great company with a bright future and with competition from the Internet proving to be benign, to date. However, the current valuation, we believe, is stretched and needs to come in. Now may be a good time to take chips off the table and come back in at better entry points, below $1.40. The growth characteristics are no doubt faster for Sirius XM relative to the more traditional media companies, but not enough to warrant a multiple three times ahead.
Further, a Liberty Media (LCAPA)-driven transaction, see article here for background, is likely over a year and half away. (As an aside, I found the 300 or so comments in response to this article particularly insightful.) While we expect the company to continue to perform on subscriber growth, earnings, and free cash flow, the anticipated Liberty Media transaction is likely the only major catalyst for this stock.
4Q earnings are scheduled for February 15, 2011 prior to the market open. Consensus estimates can be seen here. We had noted our concern that management had not pre-announced 4Q, but our concerns have now lessened. Results are likely to come in within expectations and 2011 guidance on subs and EBITDA, if they provide it, should be within the current Street range. Therefore, we are not expecting a major run in the stock either way post earnings.
No doubt we will catch enormous heat for this view from the Sirius XM lovers but valuations ultimately matter.
So lets take a deeper dive into the drivers of this model to see if we are missing something fundamental that supports this high multiple. The key drivers are subscribers, churn, ARPU, subscriber acquisition costs (SAC), programming expenses, auto installations, retail, capital expenditures, and free cash flow.
Subscribers / Auto installs / Retail. Sirius gets a sub when a subscription is sold either at retail and when a customer calls in to activate their subscription or through an install when an auto is sold. Installs in used cars work similar to retail. Sirius has factory install relationships with several auto manufacturers, where the radios are installed on the assembly line. Factory installs can account from anywhere from 25% to 100% of an OEM vehicle’s fleet, depending on the OEM. There are also relationships where the radio can be installed through a dealer.
With some OEMs such as GM (NYSE:GM) , Honda (NYSE:HMC), Ford (NYSE:F), and Chrysler (DCX), radios are factory installed and buyers are offered promotional periods where they get SIRI service for free for three months to a year depending on the OEM. SIRI counts these subs as promotional subs. These subs have averaged over 3 million over the past few years or at a high teens percentage of the total sub mix. Now SIRI, in some cases but not all, does get paid by the OEM for a few months of each promotional period so it is all not lost revenue to them. In some cases, SIRI counts subs when a car has not been sold, aptly named parking lot subs. Those are a small percentage of the mix, but nonetheless, have the effect of reducing reported SAC per gross add.
Conversion rates of promotional subs have averaged in the high 40% range over the years, proving that the factory install deals are extremely effective at driving subscriber growth. Hence, as long as consumers purchase cars, SIRI is assured a steady stream of Subs. This is huge positive for the model.
On the retail side, trends have weakened over the years as consumers are less likely to purchase satellite radios from consumer electronics retailers. Most analysts see retail subs declining into perpetuity. This is a negative.
Now churn rates have averaged about 2% over the years and is not dissimilar to traditional sub models like satellite and cable TV. So nothing to fret here and in fact churn held up well during the recession. This is more a testament to the viability and longevity of low priced content during depressed economic times. This is a positive. Now competition from the Internet, if it becomes real, can impact churn. But in our view, Internet models cannot afford big checks for exclusive content like Howard Stern, Oprah, Martha Stewart, and sports, etc., so churn should not be an issue. But one never knows .
So with factory installs supporting growth, stable churn, partially offset by declining retail activations, the model works extremely well. If the factory install model breaks, SIRI’s model breaks down, but that is unlikely from where we stand, but anything can happen. With all that laid out, we see nothing here to warrant a multiple 3x that of traditional media sub models. Correct us if we are wrong.
Programming Expenses. The good thing here is that SIRI has been able to renegotiate down their programming expenses, re-signing their talent at much lower rates. Kudos to the CFO for his discipline. This is great but will not hold under a competitive environment if one surfaces. But for now the model is safe here and EBITDA and EBITDA margins should continue to improve. But doing a good job at managing expenses should not warrant a multiple 3x its peers. Again correct us if we are wrong.
Revenue shares. Admittedly, we do not have much insight here other than to say that we believe that SIRI can reduce the revenue share to GM over the next few years. This is a positive for margins and free cash flow but we cannot confirm this.
ARPU / SAC. ARPU has been coming up over the past few years and is likely to continue to trend up if SIRI succeeds in raising prices. A positive. However, there is a point where price elasticity will impact demand for the service, so there is a limit to how much SIRI can raise prices.
On SAC per gross add, the company has done an excellent job at reducing that metric from a few hundred dollars to around $60 today. However, they are unlikely to achieve more efficiency in that metric going foward, and if they aggressively persue the used car market, an huge opportunitity, subscriber acquistion costs will likely go up due to the marketing spend needed to support that initiative ... a neutral. There is nothing in those two metrics to warrant paying 3x other traditional media.
Capital Expenditures/Free cash flow. Cash is king in our view and improving free cash flow conversion is a reason to continue to buy into stocks. With cash pilling up companies can do stuff like buy back stock, pay dividends, lever down, explore new accretive markets, buy other companies that are accretive. Free cash flow conversion is the best predictor of a model’s efficiency in our view.
SIRI is through its satellite replacement cycle this year and won’t incur those capex expenses for another five years. This should lead to a significant boost to free cash flow as maintenance capex is only about $50 million or so a year. Free cash flow should jump from an estimate of $140 million in 2010 to $1.5 billion annually in five years. That is all predicated on the satellites in space functioning properly. Satellites do fail and will have to be replaced by spare satellites, which, in that event, a new spare satellite will have to be deployed, which would crimp those free cash flow assumptions. So there is risk here.
We estimate that in a perfect world for SIRI, in which consumer preference for satellite radio maintains, factory install model continues to work, SAC is manageable, and no satellites fail, SIRI can grow its free cash flow at over a 50% CAGR in five years. This is great and is reason to own the stock but this free cash flow could likely go to servicing debt rather than for share repurchase and dividends. But if that is what you want you are better off buying companies like Time Warner, who is returning about $3 billion to shareholders each year in the form of dividends and share purchases and is doing one of the best jobs at capital efficient utilization of any companies in the TMT space, in our view. The shares trade at a reasonably multiple below 8x 2011 EBITDA, and carries less risk.
So the takeaway is that the only meaningful metric that gets us going for SIRI is the free cash conversion, which by the way is the result and dependence on all the above metrics moving together in the right direction over the next years. However, given that SIRI is unlikely to return this cash to shareholders (we have not seen any comments to the contrary), and there are better more efficient models in the TMT space, we cannot justify owning the shares at the current multiple.
Hate mail welcomed. We learn from them, really.
Next up. How sustainable is the most the important part of the business model – factory installs. If that breaks, the model breaks.