Anti-Models - Charter Communications
I attended a fascinating fireside chat between Ted Seides and Randall Stutman in January of 2022. One of the most interesting takeaways I had from their on-stage conversation was the concept of anti-models. Randall has studied great leaders across many organizations and noted that it is valuable to study what worked for some, but equally valuable to study obvious negatives, i.e. the anti-models. The idea was that if we can properly identify the things that we want to capture and emulate AND the ones we definitely do not want to capture or emulate, we can reach our goals more efficiently.
Applying this to the investment process, we can learn a lot by looking for companies that meet our quality criteria, but we can also learn a lot by studying companies that do not pass, especially if we force ourselves to be rigorous in that process. Doing so helps accelerate pattern recognition which helps eliminate new ideas earlier in the research process.
This post will discuss an anti-model for Recurve’s Builder Company framework: Charter Communications (NASDAQ: CHTR). This is not a recommendation to sell or short Charter, but rather a practical example of a seemingly “good” business that we happily pass on, even though it is a core position in many “quality compounder” portfolios. Reasonable people can disagree, but we prefer to invest when we have much more conviction in the long-term range of outcomes and when competitive positioning is strengthening over time, not weakening.
Let’s dig in.
Applying The Builder Framework
Our investment framework walks through several high-level components of analysis. We seek the following characteristics in our Builder Companies:
Market leadership
Healthy end market dynamics
Validated customer relationships and unit economics from those relationships
Customer value proposition and business economics improve with scale
An owner-oriented management team
Satisfying these components is necessary but not sufficient for us to be interested. We care about other attributes as well, like valuation, conviction levels, disruption risk, and others. Charter checks the box for some of the characteristics we care about, but misses some important ones.
We’ll walk through all the components below, but spend the bulk of the time on #2 to explain why we think the end market is unhealthy.
1. Market leadership
Charter has about 55% market share in its footprint, so it is clearly the market leader in its regions. However, we also care a lot about change in market share. On this metric, Charter has been pretty stable, but it has been slightly losing share in the important broadband internet market. To move forward, we’d need to get comfortable with recent share losses being a transitory trend to move forward. Unfortunately, we are not comfortable on this front. More on this below.
2. Healthy end market dynamics
This is where Charter (and all consumer cable/telco) fails for us. In this section, we’ll provide our view on why cable companies succeeded in pursuing monopolistic strategies in the past, and why they are less likely to succeed in doing so in the future. At a high level, we believe this market is transitioning from an “unregulated monopoly” structure to a “concentrated competitive” structure.
For most of the last 15-20 years, US cable companies methodically beat up copper-based incumbent telecom companies for subscribers, offering significantly faster internet speeds (100+ mbps vs. 30 mbps or worse) high but tolerable prices. They also offered big fixed pipes into homes which delivered linear TV bundles and on-demand capabilities via their set-top boxes.
Two major trends over the last 10-15 years have disrupted the health of the end market: (1) streaming video served over the internet, and (2) 5G fixed wireless broadband. The growth of Netflix and other streaming services is well-known – we don’t need to spend much time there. For many years, cable companies explained away their video losses, claiming that the bulk of their profitability came from their broadband businesses. Video was a lower gross margin business because of its ever-increasing content costs, but broadband is very high margin. The loss in video gross profit was manageable as long as they could capture increasing value from higher prices and high retention of broadband subscribers. After all, if subscribers were consuming video over the internet, they should value and pay up for their increasingly critical high-speed internet connection. This made sense until recently.
5G opened a crack in cable’s armor. We’re going to spend some time walking through some background on how and why we believe this happened. For most of the industry’s history, the US residential connectivity market largely evolved into each market having two primary competitors: the incumbent voice provider and a competitive cable “overbuilder” (so called because they expended risk capital to “overbuild” the existing telecom footprint with their own infrastructure). Leveraging a significant infrastructure advantage, cable was able to gain significant share over time.
For decades, cable infrastructure offered a superior pipe into the home, consisting of a core fiber network terminating in homes with insulated coaxial cables (these networks are referred to as HFC, or Hybrid Fiber Coax). The insulation around coaxial cables reduces signal loss over longer distances such that HFC cable companies could deliver superior broadband speeds to more homes, and more cost-effectively, compared to the DSL networks that were adapted from voice networks with twisted pair copper last-mile drops into homes. Twisted pair copper connections suffer from extreme signal degradation over short distances, reducing fixed telecom carriers’ ability to serve high-speed broadband to many customers at a reasonable cost. The chart below shows it well – DSL degrades by 50% at 1 km, meaning fiber-fed network nodes have to be close to customers’ homes to provision high-speed services, which has been prohibitively expensive to do at scale.
Source: Frank Phillipson, “Efficient Algorithms for Infrastructure Networks: Planning Issues and Economic Impacts
To quickly summarize, cable had a huge infrastructure-derived cost and capacity advantage vs. old copper telecom networks for delivering broadband into homes. That is still true today and where most cable bulls derive their conviction, but there is new competition that has changed the market structure: high-capacity 5G networks.
While the building blocks of 5G networks, as constructed today, cannot support widespread, massive (>50%) market share gains in home internet markets, we believe they deliver enough capacity to disrupt the subscriber growth, churn, and pricing dynamics of the cable companies to make a bit of a mess of the market. Let’s dig into “how” for a moment by taking a journey through wireless spectrum and technology. This is going to be a summary that gets a bit technical, but hopefully the logical output will be clear by the end.
For some background, first-generation, high-throughput LTE networks were revolutionary for mobile applications. For the first time ever, we could consume significant video on demand, order an Uber, browse Facebook or Instagram, and more – all on the go. The upgrade from 3G to 4G unlocked this capability at scale through order-of-magnitude increases in spectral efficiency on a like-for-like basis. The evolution from 4G to 5G is more evolutionary on a like-for-like basis, but the way it’s been deployed in the US has some important nuances. Let’s walk through some of the details:
Carriers have deployed their primary capacity layers of 5G on mid-band spectrum, generally between 2 ghz and 4.5 ghz. These frequencies are considered “mid-band” and they allow carriers to use new radio technologies that couldn’t be used in low-band frequencies that are more appropriate for coverage. Carriers’ 5G coverage layers at lower frequencies behave more like upgraded LTE.
Mid-band spectrum is best used for capacity layers has some significant advantages vs. low-band spectrum used to lay the LTE foundation. In particular:
a. Carriers have bigger blocks of contiguous spectrum in mid-band, increasing total capacity and also reducing the allocation of guard bands and increasing the total usable spectrum per large block license.
b. Higher frequency = shorter wavelengths. This is both good and bad. It’s bad for signal propagation and coverage from a single site, but good for using new technologies like massive MIMO and beamforming, which allow for a 5-10x in spectral efficiency. More on this point later.
c. Uplink and downlink transmissions are allocated using time division duplexing (TDD) rather than frequency division duplexing (FDD) which is how spectrum was allocated when voice traffic was symmetrical. In TDD, the ratio of downlink transmissions (time frames) can be modified to more closely match real-world uplink/downlink traffic patterns, whereas in FDD, half the spectrum is dedicated on a fixed basis to uplink and half is dedicated to downlink. Because downlink tends to be 90% of mobile data traffic, TDD is more efficient.
As mentioned above, Massive MIMO antenna arrays created the most significant capacity gain on a like-for-like basis, increasing efficiency by an order-of-magnitude. These antenna arrays require transceivers to separated by half a wavelength. For T-Mobile’s 2.5 ghz, they must be separated by 6cm. In C-band where AT&T, Verizon, and T-Mobile have deployed 5G services, half a wavelength is about 3-4 cm, meaning they could have up to 2x the antennas as a T-Mobile radio in the same-sized antenna box. Antenna boxes must be small enough to not cause wind shear problems on towers, so this is an important limiting factor.
The higher the frequency band, the shorter the wavelength. The shorter the wavelength, the more antennas a carrier can install into a Massive MIMO radio. The chart below shows why this is important: the more antennas in an array, the more spectrally efficient that radio will be, and the more capacity that radio can generate.
Source: Spectral Efficiency Analysis in Massive MIMO using FBMC-OQAM Modulation
To make this more practical, let’s apply some back-of-the-envelope math to illustrate:
· Early LTE networks were deployed on 20 or 40 mhz of FDD spectrum at 700 mhz frequencies. We’ll use 20 mhz for demonstrative purposes (10 mhz up x 10 mhz down). 20 mhz of LTE at 700 mhz generated about 4 bits/hz of spectral efficiency. Downlink capacity = 5 bits/hz x 10 mhz = 40 mbps from an antenna.
· Modern 5G massive MIMO antennas with 256 transceiver components using 256 QAM (black dashed line above) can do about 40 bits/hz. Verizon, for example, has an average of 160 mhz of C-band spectrum nationwide. 80% of this can be positioned to downlink capacity. Its downlink capacity is equal to 40 bits/hz x 160 mhz x 80% downlink = 5,120 mbps.
From the back-of-the-envelope math shown above, we can see that Verizon’s mid-band 5G, deployed as described above, can generate up to a 128x improvement in capacity vs. a 20 mhz block of first-generation 4G LTE. The gain is decomposed into a 10x increase from massive MIMO spectral efficiency gain (4 bits/hz in LTE to 40 bits/hz in 5G), and the remainder from utilizing a bigger block of spectrum (160 mhz vs. 20 mhz) with a more optimal skew to downlink traffic (128 mhz downlink vs. 10 mhz downlink). Apply this capacity gain to carriers’ mid-band coverage areas since C-band and 2.5 ghz were deployed, and you get a rough idea of how much capacity has been released into the market.
Why go through this technical explanation? Because between AT&T, Verizon, and T-Mobile, there has been significant new mid-band spectrum deployed over the last three years, all of which is capable of delivering these massive increases in capacity in their wireless networks. Each carrier has different spectrum blocks and frequencies, different cell site topology, and different coverage and capacity strategies, but all have been using the same new technologies to extract material gains in their networks. With a step-change in capacity unlocked by these new technologies, they’ve created excess capacity which can be monetized by selling home internet services. While this is overly generalized, we would analogize it to the chart below, whereby mobile demand compounds steadily but the 5G era began with a significant step-up in supply growth:
Source: Recurve Capital LLC Estimates
With new technologies creating significant excess capacity and faster speeds across their networks, wireless carriers, for the first time, have begun to serve 5G fixed wireless customers with speeds that are competitive with cable companies. They do not have unlimited capacity like a fiber network, but collectively, they believe they can serve about 15-20 million homes over the medium-term, with the flexibility to turn away interested customers if their networks are stressed and it’s uneconomic to increase capacity. This is very different from a traditional overbuilder, which, per its concession agreements, typically must cover and must offer service across its entire footprint, regardless of economic viability. Over time, they also can upgrade and optimize their last-mile connections with different technologies (more mid-band and/or mmWave spectrum, fiber, etc.) to evolve the way they serve customers.
The wireless carriers bundle their fixed broadband with wireless plans at disruptive prices. I pay $30/month for T-Mobile Home Internet and I’m often able to achieve 500 mbps of download speed. This compares to the $80/month I was paying Spectrum (Charter) for 100 mbps. The prior predictable model of cable companies serving customers with gradually improving speeds in exchange for regular price increases, year in and year out, may need to change to compete with these new competitors.
To wrap up this section, we think end markets for cable companies used to be healthy and supportive of monopolistic behavior, but they aren’t anymore. Reasonable people may disagree with our take on this because the mobile carriers do not have infinite supply of wireless capacity to serve homes and they get much higher yield per mbps by selling mobility services, but we think these big national brands, armed with excess capacity and competing at disruptive prices, likely degrade the economics for incumbent monopolists, regardless of wireless carriers’ ultimate ability to serve all customers in the market. It begs the question: how much of a market has to be priced disruptively to impair the economics for the rest of the market? Our view is that if mobile carriers can address 20% of households for internet service but speak to 100% of them because of their brand and distribution reach, it’s enough to threaten and reduce economic returns across the board. The important point for us is that the number of brands competing for home broadband is materially larger than before, and the infrastructure supporting those brands is sufficient enough to cause damage to the former monopolists in the near- to medium-term, and leave enough open questions for us over the long-term to stay away.
In concentrated competitive markets, especially in mature industries, typically share gains and losses are highly visible to each competitor. Management teams must answer how and why they are gaining or losing share and formulate new strategies to improve their performance. These markets tend to see strong price competition and marginal value leaking to consumers, away from equity holders.
Of course, over time wireless carriers also could step back from home internet services as they run out of excess capacity, returning the market to a more favorable oligopoly-like competitive structure. However, wishing for an outcome doesn’t make it likely, and our conclusions after running through the analysis above is enough for us to pass and put this in the “too hard” pile.
3. Validated customer relationships and unit economics
Our primary goal in this stage of our analysis is to ensure the company in question hasn’t subsidized growth to such a degree that we can’t be sure customers love the product or service at prices that would produce attractive levels of profitability. In other words, do we know what unit economics look like when prices are normal and customers are happy?
In this case, Charter is a mature company with longstanding customer relationships and rather stable unit economics, although we believe they are becoming less stable due to the growing competitive disruptions discussed above. We also think Charter and other cable companies have flexed their monopoly power and that customers are well aware of the imbalance of power in the market. Without the new 5G entrants, we’d judge this as favorable to Charter. With new broadband competition, we tilt to an unfavorable view from worsening unit economics.
4. Customer value proposition and business economics improve with scale
Cable companies carry high fixed price/low variable price cost structures, which we tend to like. They have attractive recurring subscription revenue streams and, for most of their existence, have been able to raise prices faster than their costs. Both are good dynamics, generally speaking.
We do have a problem with the value proposition to customers on a couple dimensions, which we will walk through below.
Telecom companies enjoy deflationary forces on their costs per MB delivered to customers over time which is supportive of healthy progression in price/cost spreads. Most cable companies have pursued a “more for more” pricing strategy which we tend to like. $50 for 50 mbps became $60 for 100 mbps and $75 for 1 gbps – more for more, always reducing the price per mbps. It was a successful recipe for a long time. That said, the customer service attributes of their businesses outside of “more for more” broadband pricing have been terrible. Specifically:
1. Both content and distribution companies were so focused on squeezing every drop out of linear video that they got disrupted by streaming video, initially populated by library (not exclusive) content distributed by Netflix, i.e. the value proposition that won consumers over was low-cost, on-demand library content over the internet. Had the ecosystem been more focused on satisfying customers and delivering content how and when customers wanted it, perhaps the incumbents could have had a bigger influence on the new video market structure and enjoyed a better share of streaming economics, while also keeping customers happy along the way. The classic innovator’s dilemma caused significant problems for the incumbent content and distribution companies, and they should have managed this better.
2. Cable companies are notorious for their horrendous and frustrating service with regard to provisioning and installing services, resolving service issues, and disconnecting service. These business practices pushed most customers to accept their cable services begrudgingly, with most people eager to try viable alternatives if they were to come along. With no great alternatives, customers stuck around. Low NPS scores and bad blood among customers opened the door for fixed wireless 5G to gain share quickly with a service that can be self-installed easily and has much higher NPS scores.
Charter’s old model of pursuing monopolistic pricing and service quality is the antithesis of what we look for in our Builder Companies. We want customers to be thrilled when they use our companies’ products and services, not desperate to jump ship as quickly as possible.
5. Owner-oriented management team
This may be a controversial take given John Malone’s involvement, but we believe Charter talks the talk of an owner-oriented management team, but does not walk the walk. Their orientation to FCF/share is the right one, but their capital allocation strategy of using levered buybacks regardless of valuation seems to have done more to enrich the managers than reward long-term investors, especially in light of the competitive disruptions which were exposed by management’s short-term and medium-term focus while implementing monopolistic strategies. The company has spent over $70 billion on buybacks at a weighted-average price over $450/share. It’s an awful track record for a company that now has a highly leveraged balance sheet as an at-risk incumbent in a mature industry that’s facing new competitive disruptions.
We believe a stronger owner-oriented management team would have struck a better balance between opportunistic capital returns, balance sheet health, and investing in product innovation and customer service to maintain healthier relationships with customers. We see signs that the management team is more interested in maximizing the time they get to spend in their roles, moving slowly in an industry with accelerating innovation, rather than maximizing equity value.
Closing
We have spent many years studying cable and telecom markets, and currently we believe US consumer connectivity markets face far too much competitive disruption risk for our liking. Over the years, we have owned various companies in this market, including Charter and T-Mobile, but are not involved anymore. Both wireless and fixed connectivity markets are seeing new competition - fixed entering wireless and wireless entering fixed - from big existing brands with significant advertising and distribution reach, and bundling/discount-oriented pricing strategies that have proven to be disruptive. No one is safe and we have little confidence in our ability to forecast future competitive dynamics and earnings power for each company within a reasonable range of outcomes.
An important part of our process is to walk through anti-models of companies that do not satisfy our foundational pillars for Builder Companies. This work helps us refine our idea selection process, our research and due diligence process, and improve our filters, all of which raises the odds that the Builder Companies we find and own have passed a series of rigorous tests.
