The Centurylink Conundrum

It’s been over 4 months since I made my overtly positive 2018 EBITDA call on Centurylink (CTL), as per this February post. I estimated EBITDA guidance for 2018 around $9.25 billion whereas CTL’s actual guidance was between $8.75 billion to $8.95 billion. In my previous CTL post back in August 2017, the base case 2018 EBITDA was much closer to the mark at $8.95 billion! The reason for my $0.4 billion overshoot was an over optimistic reaction to the comments CFO Sunit Patel made earlier in the year on a possible 5%-7% margin improvement over the next 3 to 5 years. In my defense, the exhibit below comparing the cumulative margin improvement following the LVLT/TWTC merger compared to those now forecast by analysts from the CTL/LVLT merger shows how one could see more upside than indicated by guidance from the new management at CTL (i.e. the old LVLT management team after their successful reverse takeover of CTL).

click to enlarge

However, the LVLT/TWTC merger was a very different deal than the CTL/LVLT one. For starters, TWTC was a growing fiber based business, at both revenue and EBITDA lines, when it merged with LVLT whereas CTL is a declining one, at both revenue and EBITDA lines, with over 40% of its standalone business in legacy services.

I have rebuilt my model on the combined entity and carefully considered the top and bottom line impact of the declining legacy business on the combined CTL/LVLT projections in addition to the potential cost savings compared to those articulated by CFO Sunit Patel and new CEO Jeff Storey (I am assuming $1 billion of operating synergies compared to the guided figure of $0.85 billion). The graph below illustrates that looking at historical proforma margins on a combined business to project the future is misleading given the underlying trends at CTL, particularly the declining legacy business, and the improving margins at LVLT from the TWTC synergies.

click to enlarge

For the newly combined business, I estimate the old legacy business will make up approximately 25% of the new CTL’s revenue base. Historically, for the combined business, I estimate the legacy business (split 65:35 between the enterprise & consumer businesses) has been declining on average quarter on quarter by 2.5% over the past 6 quarters whereas the strategic (i.e. non-legacy) business has grown 0.5% on average. The strategic growth rate has been lower on average in recent quarters if Q1 2018 is excluded. The previous quarters poor performance could be due to the uncertainty over the merger, but the performance of the key strategic business will be an important metric to watch in future quarters (unfortunately they do not split the business out this way in their reports anymore). The forward quarter on quarter decline/growth rates for the legacy/strategic blocks are critical in determining future revenues and margins. The graph below shows the different impacts on annual revenue growth and EBITDA margins for different sets of legacy and strategic quarterly declines/increases.

click to enlarge

This analysis shows that the underlying business is facing the headwind of up to a 3% revenue decline in 2019 and up to 2.5% decline by 2022. As a result, underlying EBITDA margins also could be facing an annual 40 basis point decline from 2019 through 2022. This decline in underlying EBITDA margins explains an element of the difference in cumulative EBITDA margin improvement in the first graph above.

For my base scenario, I have assumed a quarterly revenue decline of 2.75% on legacy business (slightly worse that the past 6 quarter average of 2.5%) and a quarterly increase of 0.50% in strategic revenue (in line with the past 6 quarter average). I have also included an increase in EBITDA margin due to new business that the enlarged group can attract due to its larger footprint and relevance (the revenue impact is minimal as it will likely mainly be from existing clients scaling up although the margin impact could be more significant). I have sense checked the resulting revenue figures against independent projections using the new business classifications presented by the firm. The breakdown of the different cumulative impacts in my base scenario are shown below.

click to enlarge

These figures get to just within the 5%-7% range of margin improvement articulated by Sunit Patel with 5.3% combined improvement after the 3rd year, leaving room for further margin improvement in subsequent years. The graph below shows my base scenario revenue figures using the new classifications from CTL.

click to enlarge

This time around, I will not be considering an optimistic scenario but rather focusing on the downside to my base selections. Upside to my base scenarios is possible if recent revenue trends improve, as a result say of rapid 5G deployments (see this post). Upside could also come from CTL’s deep enterprise network being tempting to possible acquirers in a vertical M&A frenzy, although that sounds a bit like wishful thinking to my mind. The balance of probabilities is more likely to be on the downside in terms of revenue (although I do have confidence in CTL’s management ability to manage the various levers to hit their EBITDA targets).

For my pessimistic selections, I have assumed an accelerated quarterly revenue decline of 3% on legacy business and a flatlining quarterly change of 0% in strategic revenue (due to pressure on the enterprise business from increased software enabled competitors). These are fairly brutal assumptions. The impact of new business, particularly its impact on margins, is also assumed to be diminished compared to my base assumptions. Again, I sense checked these top-line figures by projections using the new business classifications from CTL, as below. These projections clearly show a business model under significant pressure.

click to enlarge

Taking all of the above factors into account, my revenue and post synergy EBITDA projections come out as per the graph below.

click to enlarge

On CTL’s debt, despite the issues surrounding LIBOR as a base rate for the floating debt (see this post), I am reasonably comfortable given approx. 65% of the debt is fixed. The debt load is high, at net debt to mid-point 2018 guided EBITDA of 4.2, but manageable (one area where Sunit has proven his ability is debt management!). I estimate that for every 25 basis point increase in the base floating rate (there are alternatives to LIBOR detailed in its floating credit facilities which is important given LIBOR’s likely replacement by something like SOFR, as per this article) the impact on CTL’s total debt interest rate is 8.5 basis points. Also, any significant debt repayment is not due until 2020 which gives time to ensure operating efficiencies are delivered. Of course, if the overall business model is in trouble despite achieving operational targets (e.g. software based telecom serves disrupt CTL) as per the pessimistic scenario, the debt load will become a big issue by 2020 (I still have the scars from the telecom bust)!

The key question concerning CTL in the short term is the sustainability of its dividend, given its current dividend yield of over 11.5% is amongst the highest in the S&P500. Under my base scenario, I project that the dividend is sustainable, just. It will be tight particularly when CTL will want to be demonstrate continued progress on deleveraging year on year. Under my pessimistic scenario, I assume a 50% dividend cut in 2019 would be required.

My valuation for CTL under my base and pessimistic scenarios is $18 and $10 respectively. Given the stock currently trades around $18, the market is indicating a belief in the new management team and its guidance. I have a high degree of confidence in current management and their ability to navigate the integration of CTL/LVLT and the challenges ahead in telecomland. However, the dividend sustainability issue will not be resolved until we see the quarterly progress over the next few quarters. Well into 2019 would be my guess.

It’s also likely that markets will be increasingly volatile over the next 12 months (eh, the market seems to be waking up to the folly of Mr Trump’s trade war this morning). Highly indebted firms will likely be battered as we get deeper into a tighter monetary environment, particularly those without topline growth. Any dividend cut will hit the stock heavily. I would not be surprised if the stock fell as low as $6 following a 50% cut. Given the juicy dividend yield will provide the cashflow, buying options seem a sensible means of protecting against such a downside. Ultimately, I have high hopes for CTL but this one is not for the faint hearted. It’s a risky stock, as the dividend yield implies, in an increasingly unbalanced market and it has to execute flawlessly over the coming quarters to justify the risk.

5G: Telecom Hype or Saviour

John Legere of T-Mobile is a canny operator and knows how to play the sycophant to Trump’s nationalist instincts in touting the ability of a combined T-Mobile & Sprint to invest in a super-charged 5G roll out, as per this presentation, playing the job creation and beat the Chinese technological advancement cards. Legere cites an Analysys Mason report commissioned by the US industry lobby group CTIA to back up such claims which in turn cites an Accenture report from 2017 on 5G in the US which claims that “telecom operators are expected to invest approximately $275 billion in infrastructure, which could create up to 3 million jobs and boost GDP by $500 billion”. In 2016, the European Commission in this report stated that 5G “investments of approximately €56.6 billion will be likely to create 2.3 million jobs in Europe”. An IHS Markit 2017 report commissioned by Qualcomm claims that in 2035, “5G will enable $12.3 trillion of global economic output” and “the global 5G value chain will generate $3.5 trillion in output and support 22 million jobs” on the basis that “the global 5G value chain will invest an average of $200 billion annually”.

These are fantastical figures. Many assumptions go into their computation including the availability, range and cost of spectrum plus infrastructure spend and policy in relation to streamlining procedures and fee structures for the deployment of the small shoe-box cell sites (between 10 to 100 more antenna are required for 5G than current networks). Larger issues such as privacy and security also need to be addressed before we enter a world of ubiquitous ultra-reliable low latency networks as envisaged by the reports referenced above. Those of us who lived through, and barely survived, the telecom boom of the late 1990s can be forgiven for having a jaundice view of a new technology saving the telecom industry. This blog illustrates some of the challenges facing the wired telecom sector and the graph below shows the pressures that the US mobile players are under in terms of recent trends in service revenues.

click to enlarge

The mobile service revenues trends are remarkably similar to those in the enterprise and wholesale space. The graph above also shows the rationale for the T-Mobile/Sprint merger in terms of size as well as the impact of T-Mobile’s aggressive pricing strategy. All these trends are in the context of the insatiable increase in bandwidth traffic, as illustrated by the IP figures from Cisco below.

click to enlarge

This report from 2017 by Oliver Wyman is one of the better ones and contains some illuminating context for the 5G era. It shows that in Europe despite a 40% annual increase in mobile subscribers and a 36% annual increase in European IP traffic from 2006 to 2016, mobile service revenue and total telecom service revenue decreased by 22% and 19% respectively, as per the graphic below.

click to enlarge

The Oliver Wyman report concludes as follows:

“In the next five to ten years, demand in fixed-line broadband bandwidth will grow exponentially, leading to speeds that can only be supplied by FTTH/B. Mobile broadband demand will follow in parallel. Virtual reality is the “killer app” that will drive massive demand. Mobile broadband supply will begin to reach its limits, with spectral efficiency gains and additional attractive spectrum in the current bands not growing as fast as they have in the past. High-frequency beam technology in 5G will be radically new and will be able to meet future demand. At the same time, however, it will create massive mobile backhaul demand. The outcome is likely to shake the industry, leading not only to a new balance of power between mobile-only and integrated/fixed-line operators, but also to new potential revenue growth for the first time in many years.”

Another interesting graphic from the report, as below, is the historical and projected broadband usage.

click to enlarge

This 2017 report from Deloitte argues that “5G, across both the core and radio access network, stands to have a potentially greater impact on the overall ecosystem than any previous wireless generation”. Deloitte sees a “convergence of supply between wireline and wireless broadband, as almost all devices become connected over short-range wireless”. Deloitte concludes that “with an increasingly converged ecosystem of network and content players, an increasingly software-managed and defined physical networking space, and the demands and needs of consumers becoming complex enough that they no longer can manage individually, 5G and its associated technologies may have the power to reset the wireless landscape”.

This paper from an Infinera executive called Jon Baldry highlights the need for “improvements to the overall network infrastructure in terms of performance, features and bandwidth” to support 5G “using software-defined networking (SDN) control and network functions virtualization (NFV) will play a major role in the optimization of the network”. I came across an interesting claim that SDN and network virtualisation can reduce opex and capex by 63% and 68% respectively compared to traditional telecom networking. Baldry concludes that “these improvements will drive new fiber builds, and fiber upgrades to an ever-growing number of cell sites, creating significant opportunity for cable MSOs and other wholesale operators to capture significant share of cell backhaul and fronthaul services for 4G and 5G mobile networks”.

Whether all these investments and resulting new networks will halt the declining revenue trend for the telecom sector or merely provide a survival avenue for certain telecoms is something I have yet to be convinced about. One thing seems certain however and that is that tradition telecom models will change beyond recognition in the forthcoming 5G era.

Bumpy Road

After referring to last year’s report in the previous post, the latest IMF Global Financial Stability Report called “A Bumpy Road Ahead” was released yesterday. Nothing earth-shattering from the report when compared to previous and current commentary. The following statements are typical:

“Many markets still have stretched valuations and may experience bouts of volatility in the period ahead, in the context of continued monetary policy normalization in some advanced countries. Investors and policymakers should be cognizant of the risks associated with rising interest rates after years of very easy financial conditions and take active steps to reduce these risks.”

and

“Valuations of risky assets are still stretched, with some late-stage credit cycle dynamics emerging, reminiscent of the pre-crisis period. This makes markets exposed to a sharp tightening in financial conditions, which could lead to a sudden unwinding of risk premiums and a repricing of risky assets. Moreover, liquidity mismatches and the use of financial leverage to boost returns could amplify the impact of asset price moves on the financial system.”

With the US 10 year breaking 2.9% today and concerns about a flattening yield curve, the IMF puts global debt at $164 trillion or 225% of GDP (obviously a different basis from the IIF’s estimate of global debt at $237 trillion or 318% of GDP) and warns about the US projected debt increase due to its “pro-cyclical policy actions”.

In a chapter the IMF calls “The Riskiness of Credit Allocations” it presents an interesting graph, as below, using its financial conditions index which uses multiple inputs constructed using a methodology that’s wonderfully econometrically complex, as is the IMF way.

click to enlarge

The IMF warn that “a variety of indicators point to vulnerabilities from financial leverage, a deterioration in underwriting standards, and ever more pronounced reaching for yield behaviour by investors in corporate and sovereign debt markets around the world”.

Nero fiddles

This week it’s Syria and Russia, last week it was China. Serious in their own right as these issues are, Donald Trump’s erratic approach to off the cuff policy development is exhausting markets. In the last 60 trading days, the S&P500 has had 9 days over 1% and 12 days below -1%. For above 0.5% and more than -0.5%, the number of days is 21 and 15 respectively! According to an off the record White House insider, “a decision or statement is made by the president, and then the principals come in and tell him we can’t do it” and “when that fails, we reverse engineer a policy process to match whatever the president said”.  We live in some messed up world!

As per this post, the mounting QE withdrawals by Central Banks is having its impact on increased volatility. Credit Suisse’s CEO, Tidjane Thiam, this week said, “the tensions are showing and it’s very hard to imagine where you can get out of a scenario of prolonged extraordinary measures without some kind of, I always use the word ‘trauma’”.

Fortune had an insightful article on the US debt issue last month where they concluded that something has to give. According to an Institute of International Finance report, global debt reached a record $237 trillion in 2017, more than 317% of global GDP with the developed world higher around 380%. According to the Monthly Treasury Statement just released, the US fiscal deficit is on track for the fiscal years (Q4 to Q3 of calendar year) 2018 and 2019 to be $833 billion and $984 billion compared to $666 billion in 2017.

This week also marks the publication of the Congressional Budget Office’s fiscal projections for the US after considering the impact of the Trump tax cuts. The graphs below from the report illustrate the impact they estimate, with the fiscal deficits higher by $1.5 trillion over 10 years. It’s important to note that these estimates assume a relatively benign economic environment over the next 10 years. No recession, for example, over the next 10 years, as assumed by the CBO, would mean a period of nearly 20 years without one! That’s not likely!

The first graph below shows some of the macro-economic assumptions in the CBO report, the second showing the aging profile in the US which determines participation rates in the economy and limits its potential, with the following graphs showing the fiscal estimates.

click to enlarge click to enlarge

click to enlarge

click to enlarge

click to enlarge

The respected author Satyajit Das highlighted in this article how swelling levels of debt will amplify the effect of any rate rises, with higher rates having the following impacts:

  • Increase credit risk. LIBOR has already risen, as per this post, and large sways of corporate debt is driven by LIBOR. This post shows some of debt levels in S&P500 firms, as per the IMF Global Financial Stability report from last April and the graph below tells its own tale.
  • Generate large mark-to-market losses on existing debt holdings. A 1% increase is estimated to impact US government debt by $2 trillion globally.
  • Drive investors away from risky assets such as equity, decimating the now quaint so-called TINA trade (“there is no alternative”).
  • Divert cash to servicing debt, further dampening economic activity and business investment.
  • Restrict the ability of governments to deploy fiscal stimulus.

click to enlarge

Back in the land of Nero, or Trump in our story, his new talking head in chief, Larry Kudlow, recently said the White House would propose a “rescission bill” to strip out $120 billion from nondefense discretionary spending. Getting that one past either the Senate or the House ahead of the November midterm elections is fanciful and just not probable after the elections. So that’s what the Nero of our time is planning in response to our hypothetical Rome burning exasperated by his reckless fiscal policies (and hopefully there wouldn’t be any unjustified actual burning as a result of his ill thought out foreign policies over the coming days and weeks).

Generational Music

People of my generation, like those before us no doubt, like to moan on about the quality of modern music. When I look back at the diversity of the music from the 1980’s that I grew up listening to, I cannot but help feel that this generation is missing out.

As it happens, the 2018 Global Music Report from IFPI indicated that the multi-year decline in global music revenues has bottomed out. The 2017 industry revenues grew by 8% over 2016, with streaming revenues up 41%. This represents three consecutive years of growth after many years of decline. The music sector is one of the earliest examples of the awesome creative destructive ability of the digital revolution, as the graph below shows.

click to enlargeGlobal Music Industry Revenues 1972 to 2020 April 2018

The physical, digital and streaming revenues are obvious (e.g. CDs & vinyl, downloads & streaming). Performance rights includes the revenues generated by the use of recorded music by broadcasters and public venues. Synchronisation revenues include the revenues from the use of music in advertising, film, games and television programmes.

As a regular theme of this blog is the impact of the digital revolution under way on so many industries and the need for sectors to adapt through digital transformation of their business models, the graph above is both thought provoking and scary.

Listening to the investment pitch by Spotify this week over the future of the sector, I can’t but help think that the democratisation and disintermediation promised by the internet age has resulted, for the music sector at least, in dominant players dictating homogeneous tastes and culture. The death of individualism seems to be the result, at least until this or future generations get fed up with it.