In the past week or so I’ve seen several articles that remind me how important the Quello Center’s empirically-grounded study of net neutrality impacts is for clarifying what these impacts will be—especially since net neutrality is one of those policy topics where arguments are often driven by ideology and/or competing financial interests.
As far as I can tell, this series of articles began with an August 25 piece written by economist Hal Singer and published by Forbes under the following headline: Does The Tumble In Broadband Investment Spell Doom For The FCC’s Open Internet Order? Per his Forbes bio, Singer is a principal at Economists Incorporated, a senior fellow at the Progressive Policy Institute, and an adjunct professor at Georgetown University’s McDonough School of Business.
Singer’s piece was followed roughly a week later by two op-ed pieces published on the American Enterprise Institute’s web site. The title of the first AEI piece, authored by Mark Jamison, was Title II’s real-world impact on broadband investment. This was followed a day later by Bronwyn Howell’s commentary Title II is hurting investment. How will – and should – the FCC respond?
What struck me about this series of op-ed pieces published by economists and organizations whose theoretical models and policy preferences appear to favor unregulated market structures was that their claims that “Title II is hurting investment” were all empirically anchored in Singer’s references to declines in ISP capital spending during the first half of 2015. As a member of the Quello Center’s research team studying the impacts of net neutrality, I was intrigued, and eager to dig into the CapEx data and understand its significance.
While my digging has only begun, what I found reminded me how much the communication policy community needs the kind of fact-based, impartial and in-depth empirical analysis the Quello Center has embarked upon, and how risky it is to rely on the kind of ideologically-driven analysis that too often dominates public policy debates, especially on contentious issues like net neutrality.
My point here is not to argue that there are clear signs that Title II will increase ISP investment, but rather that claims by Singer and others that there are already signs that it is hurting investment are not only premature, but also based on an incomplete reading of evidence that can be uncovered by careful and unbiased review of publicly available information.
I hope to have more to say on this topic in future posts, but will make a few points here.
The crux of Singer’s argument is based on his observation that capital spending had declined fairly dramatically for a number of major ISPs during the first half of 2015, dragging down the entire sector’s spending for that period (though its not clear from the article, my sense is that Singer’s reference to “all” wireline ISPs refers to the industry’s larger players and says nothing about investment by smaller companies and the growing ranks of publicly and privately owned FTTH-based competitors). He then briefly reviews and dismisses potential alternative explanations for these declines, concluding that their only other logical cause is ISPs’ response to the FCC’s Open Internet Order (bolding is mine):
AT&T’s capital expenditure (capex) was down 29 percent in the first half of 2015 compared to the first half of 2014. Charter’s capex was down by the same percentage. Cablevision’s and Verizon’s capex were down ten and four percent, respectively. CenturyLink’s capex was down nine percent. (Update: The average decline across all wireline ISPs was 12 percent. Including wireless ISPs Sprint and T-Mobile in the sample reduces the average decline to eight percent.)..
This capital flight is remarkable considering there have been only two occasions in the history of the broadband industry when capex declined relative to the prior year: In 2001, after the dot.com meltdown, and in 2009, after the Great Recession. In every other year save 2015, broadband capex has climbed, as ISPs—like hamsters on a wheel—were forced to upgrade their networks to prevent customers from switching to rivals offering faster connections.
What changed in early 2015 besides the FCC’s Open Internet Order that can explain the ISP capex tumble? GDP grew in both the first and second quarters of 2015. Broadband capital intensity—defined as the ratio of ISP capex to revenues—decreased over the period, ruling out the possibility that falling revenues were to blame. Although cord cutting is on the rise, pay TV revenue is still growing, and the closest substitute to cable TV is broadband video. Absent compelling alternatives, the FCC’s Order is the best explanation for the capex meltdown.
I haven’t had a chance to carefully review the financial statements and related earnings material of all the companies cited by Singer, but did take a quick look at this material for AT&T and Charter since, as he notes, they experienced by far the largest percentage drop in spending. What I found doesn’t strike me as supporting his conclusion that the decline was network neutrality-driven. Instead, in both cases it seems to pretty clearly reflect the end of major investment projects by both companies and related industry trends that seem to have nothing to do with the FCC’s Open Internet order.
My perspective on this is based on statements made by company officials during their second quarter 2015 earnings calls, as well as capex-related data in their financial reporting.
During AT&T’s earnings call, a Wall Street analyst asked the following question: “[T]he $18 billion in CapEx this year implies a nice downtick in the U.S. spending, what’s driving that? Are you finding that you just don’t need to spend it or are you sort of pushing that out to next year?” In his response to the question, John Stephens, the company’s CFO, made no mention of network neutrality or FCC policy decisions. Instead he explained where the company was in terms of key wireless and wireline strategic network investment cycles (bolding is mine):
Well, I think a couple of things. And the simplest thing is to say [is that the] network team did a great job in getting the work done and we’ve got 300, nearly 310 million POPs with LTE right now. And we are putting our spectrum to use as opposed to building towers. And so that aspect of it is just a utilization of spectrum we own and capabilities we have that don’t require as much CapEx. Secondly, the 57 million IP broadband and what is now approximately 900,000 business customer locations passed with fiber. Once again, the network guys have done a great job in getting the Project VIP initiatives completed. And when they are done…the additional spend isn’t necessary, because the project has been concluded not for lack of anything, but for success.
Later on in the call, another analyst asked Stephens “[a]s you look out over the technology roadmap, like 5G coming down the pipeline, do you anticipate that we will see another period of elevated investment?”
While Stephens pointed to a potential future of moderated capital spending, he made no reference to network neutrality or FCC policy, focusing instead on the investment implications of the company’s (and the industry’s) evolution to software-defined networks.
I would tell you that’s kind of a longer term perspective. What we are seeing is our move to get this fiber deep into the network and getting LTE out deep into the wireless network and the solutions that we are finding in a software-defined network opportunity, we see a real opportunity to actually strive to bring investments, if you will, lower or more efficient from historical levels. Right now, I will tell you that this year’s investment is going to be in that $18 billion range, which is about 15%. We are certainly – we are not going to give any guidance with regard to next year or the year after. And we will give an update on this year’s guidance, if and when in our analyst conference if we get that opportunity. With that being said, I think there is a real opportunity with some of the activities are going on in software-defined networks on a longer term basis to actually bring that in capital intensity to a more modest level.
Charter’s large drop in capital spending appears to be driven by a similar “investment cycle” dynamic. During its 2Q15 earnings call, CFO Christopher Winfrey noted that Charter’s year-over-year decline in total CapEx “was driven by the completion of All-Digital during the fourth quarter of last year,” referring to the company’s migration of its channel lineup and other content to an all-digital format.
A review of the company’s earnings call and financial statements suggests that a large portion of the “All-Digital” capital spending was focused on deploying digital set-top boxes to Charter customers, resulting in a precipitous decline in the “customer premise equipment” (CPE) category of CapEx. According to Charter’s financial statements, first-half CPE-related CapEx fell by more than half, or $341 million, from $626 million to $285 million. Excluding this sharp falloff in CPE spending driven by the end of Charter’s All-Digital conversion, the remainder of the company’s capital spending was actually up 3% during the first half of 2015. And this included a 7% increase in spending on “line extensions,” which Charter defines as “network costs associated with entering new service areas.” It seems to me that, if Charter was concerned that the Commission’s Open Internet order would weaken its business model, it would be cutting rather than increasing its investment in expanding the geographic scope of its network.
To understand the significance of Charter’s spending decline, I think it’s important to note that its 29% decline in first half total CapEx was driven by a 54% decline in CPE spending, and that the company’s non-CPE investment—including line extensions—actually increased during that period. I found it odd that, even as he ignored this key dynamic for Charter, Singer seemed to dismiss the significance of Comcast’s CapEx increase during the same period by noting that it was “attributed to customer premises equipment to support [Comcast’s] X1 entertainment operating system and other cloud-based initiatives.”
I also couldn’t help notice that, in his oddly brief reference to the nation’s largest ISP, Singer ignored the fact that every category of Comcast’s capital spending increased by double-digits during the first half of 2015, including its investment in growth-focused network infrastructure, which expanded 24% from 2014 levels. Comcast’s total cable CapEx was up 18% for the first half of the year, while at Time Warner Cable, the nation’s second largest cable operator, it increased 16%.
While these increases may have nothing to do with FCC policy, they seem very difficult to reconcile with Singer’s strongly-assserted argument, especially when coupled with the above discussion of company-specific reasons for large CapEx declines for AT&T and Charter. As that discussion suggests, the reality behind aggregated industry numbers (especially when viewed through a short-term window of time) is often more complex and situation-specific than our economic models and ideologies would like it to be. This may make our research harder and messier to do at times, but certainly not less valuable. It also speaks to the value of longitudinal data collection and analysis, to better understand both short-term trends and those that only become clear over a longer term. That longitudinal component is central to the approach being taken by the Quello Center’s study of net neutrality impacts.
One last general point before closing out this post. I didn’t see any reference in Singer’s piece or the AEI-published follow-ups to spending by non-incumbent competitive providers, including municipally and privately owned fiber networks that are offering attractive combinations of speed and price in a growing number of markets around the country. While this category of spending may be far more difficult to measure than investments by large publicly-owned ISPs, it may be quite significant in relation to public policy, given its potential impact on available speeds, prices and competitive dynamics.
Expect to see more on this important topic and the Quello Center’s investigation of it in later posts, and please feel free to contribute to the discussion via comments on this and/or future posts.
