About a year and a half ago, I wrote an article on Sirius XM, and I want to revisit that article, address some of its claims, and discuss the reason why we sold our shares. If you haven't read the previous article, then I would encourage you to do so now as the rest of this article references it extensively.
First off, with respect to the company's competitive position, not much has changed - the company still has a wide and diverse range of content that people enjoy and are willing to pay for. Any weakness in its business would show up in a lower conversion rate and/or a higher churn rate. Conversion rates have continued to be strong, and churn has stayed in the same general, and low, range.
With all of that said, I am no longer so positive on the shares of the company. What made me change my mind?
- Redetermination of the significance of the churn rate.
- Lower than expected pricing power
- Faulty assumptions in the free cash flow conversion ratio
- Lower than expected leverage
Churn Rate - The churn rate is the number of paying subscribers that leaves the service every month. Sirius XM's churn rate, at around 1.8%, is low when compared to most other subscription-based businesses, and a lower level of cancellations will mean more subscribers, revenue and profit down the line. While the headline number is low, I have come to realize that Sirius XM's churn rate is not directly comparable to a company like AT&T. Much of AT&T's wireless churn consists of customers switching to, say, Verizon. On the other side, much of Verizon's churn consists of customers switching to AT&T. However, since Sirius XM doesn't have a direct competitor, their churn consists of customers stopping satellite radio altogether. This is important because it means that for any given churn number, AT&T can still have a healthy, perhaps even growing market, while Sirius XM would have a declining user base as a percentage of all enabled cars.
Pricing Power - Sirius XM has had lower pricing power than I anticipated, with average revenue per user (ARPU) increasing only 2.5% in total (i.e. 1.25%/year) over the past two years. I would have expected 2-3% growth in this metric per year. This is important not only for the financial impact of higher revenue, but also as a sign of the business' health - strong pricing power means that customers generally can't do without your products.
Free Cash Conversion - In my previous article, I took management's guidance for free cash flow conversion (defined as free cash flow as a percentage of EBITDA) as a given at roughly 55%. However, given reasonable estimates for the real cost of satellites (about $100 million/year or $20 million/year per satellite) and other capital spending, along with interest rates, and perhaps most importantly stock-based compensation, I believe the real cash flow conversion percentage should be in the 40%-45% range. This is very important because it means that for every dollar of revenue, I overestimated earnings, and therefore the value of the company, by around 25%.
Leverage - Management had guided to a debt/EBITDA ratio of around 4x, but they have subsequently said that leverage dedicated to buybacks will be closer to 3x with the ability to go up to 4x for any accretive acquisition opportunities.
While I still believe that Sirius XM is a good company, and I would be happy to own it again at the right (read: lower) price, I overestimated its potential earnings when I wrote the first article. Every adjustment to my estimates that I made above ultimately leads to lower earnings. My first estimate was that the company would have free cash flow of around $1.5 billion in 2018, but with all of these changes, my new estimate is for a little under $1 billion - a 35% drop.
Below, I'll do a quick update on my valuation. I estimated that it was worth around $4.00/share in August of 2014, which would mean that if everything panned out like I had estimated that shares would be now worth around $4.70.
Here is an updated chart with estimated revenue, EBITDA, free cash flow and debt out to 2018. The assumptions are as follows:
- Revenue growth of 7% - takes into account a bigger impact from churn and also slightly lower pricing power.
- EBITDA margins at 40% - same as previously estimated.
- FCF conversion at 42.5% - this is down from 55% and is by far the biggest change on estimated value.
- Debt at 3.25x EBITDA - down from 4x in the previous model.
I'll spend a second discussing the free cash flow conversion ratio as that is what has changed the most and also the most important number affecting value. We'll simply take the 2018 EBITDA estimate of $2.239 billion and add back expenses to get to free cash flow.
Depreciation and Amortization - I estimate around $300 - 325 million of D&A for 2018, which is up around 10-20% from 2015 and generally in line with historical expense growth.
Interest Expense - With debt at around 3.25x EBITDA, total borrowings would be about $7.3 billion. The company currently pays around 5.5% on its debt, but I'll use a range of 5.5%-7% to account for the possibility of higher interest rates. It's hard to see the interest rate going much below 5.5% but it certainly could be higher than 7% with higher rates in general.
Tax Expense - This should be around 37%, as all the company's income occurs in the US.
Capital Spending - I have taken the past three years of capital expenditures and added smoothed satellite costs to those numbers. For those unfamiliar with the company, Sirius XM has 5 satellites in orbit, but the costs for that come in chunks. Each one costs around $300 million and lasts for approximately 15 years, coming to a total cost of $100 million/year for all five. There have been no satellite costs for the last few years, so to estimate the "correct" amount of capital spending, I took those numbers and added the amortized cost ($100 million) of the satellites. That gives you capital spending as a percentage of D&A expense of 93%.
Putting it all together:
Even using all the most aggressive estimates (low D&A and a 5.5% interest rate) we still only get a free cash flow conversion ratio of 44%, so we can safely reject the 55% number. Just for comparison, using D&A expense of $325 million and a 7% interest rate gets you to a ratio of 41%.
One last note here is that just looking at reported cash from operations will artificially boost cash conversion because share-based payment expense is not included as a cost in the cash flow statement. Share-based compensation is always a cost and I therefore account for it as such.
With our estimates of revenue, EBITDA and free cash conversion, we'll do a quick valuation of the company:
Here is a quick rundown of the value drivers:
- FCF - The normalized free cash flow of the business and is taken from the previous chart.
- Debt - Cash flow available from incremental debt issuance to keep the leverage ratio at 3.25x.
- NOLs - Net operating losses that enable the company to not pay income taxes and record higher cash flow ($4.6 billion currently remaining).
- TL - The total of the columns.
- Discounted - The total amount discounted at an annual rate of 10%.
- Terminal value - 16x my free cash flow estimate for 2019.
Adding up the values gets to a valuation of around $17.4 billion, or $3.38/share. As you can see this is considerably lower than my estimate in 2014 of around $4.00, especially when taking a year and a half of growth into account.
Thoughts and Learnings
I largely view the first valuation exercise as a failure. Here are some thoughts while looking back on the numbers:
Business models matter - First and foremost, Sirius XM is still a fundamentally very good business, even if I miscalculated its earnings power. Companies all over the media landscape are struggling with declining subscribers and/or profitability, whether the medium is in streaming music, newspapers, or TV channels. Furthermore, their outlooks largely do not look bright, and combined with sometimes too high leverage, many of these companies' shares have been decimated. Even mighty Disney is down around 20% largely because of weakness, perceived or real, at its ESPN business.
Success in investing will always be about picking good businesses. Even though I greatly overestimated Sirius XM's earnings power, the stock is doing OK because I have been basically correct about the business - it is both growing and very profitable and it looks like that should continue.
2. Margin of safety - When I first wrote the article, shares were trading around 75% of my estimate of worth, meaning shares would have to go up ~35% immediately just to get back to fair value (which would be growing every year). Because of the large gap between price and my initial estimate of worth, even though I was wrong about the level of profitability in the business, the stock turned out to be trading at a decent price.
Sirius XM is certainly a very good company - it has a diverse collection of important content, it earns huge returns on tangible invested capital, revenue, earnings, and free cash flow should continue to increase nicely, and it has one of the most successful media investors ever as its controlling shareholder. However, for all of these traits, I believe that the shares are just too expensive.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: I am a registered investment adviser. However, this commentary does not constitute individualized investment advice. The opinions offered herein are not personalized recommendations to buy, sell or hold securities.