Archive for September, 2015
This one is impossible to answer in a 140-character reply tweet. So here is my 600-word response.
To paraphrase Bill Clinton, it depends on the meaning of “net neutrality.”
If net neutrality means barring any payments from edge providers to ISPs, then there may be no basis in law, including Title II. Check out the Phoenix Center’s amicus brief (pages 3 to 4) for a quick primer on why such ex ante prohibitions are not sustainable even under Title II.
In contrast, if net neutrality means preventing an ISP from discriminating on the basis of application or affiliation—a principle that some important players like Google appear to support—then we can design a remedy without relying on Title II. (I happen to support this principle too.)
The first remedy finds its basis in Section 706, and was telegraphed by the D.C. Circuit in its Verizon decision. Following the guidance of Cellco, the court signaled it would tolerate a case-by-case regime that grants room for “individualized bargaining” by the parties to a paid priority arrangement. (Alas, the FCC rejected this approach in its 2015 Open Internet Order.)
Such deals, if done in a discriminatory manner, could be challenged ex post by third parties or by the FCC, but—and this is key—the burden of proof would fall on the challenger. In particular, the paid priority arrangement would be presumed not to violate the non-discrimination standard, and the challenger would have to overcome that presumption.
What the court rejected was the opposite presumption. In particular, it ruled that the FCC’s 2010 OIO—which in footnote 229 rejected a “flat ban on discrimination” in favor of case-by-case adjudication—was tantamount to “per se common carriage” because any priority arrangement was presumed to be in violation of the OIO’s non-discrimination standard. This presumption placed the burden of proof on the parties to the agreement to prove that the arrangement was not in violation of the non-discrimination rule—a core aspect of common carriage.
This throwaway line in the 2010 OIO at paragraph 76 was the deal breaker for the D.C. Circuit: “In light of each of these concerns, as a general matter, is unlikely that pay for priority would satisfy the ‘no unreasonable discrimination’ standard.” The court replied: “If the Commission will likely bar broadband providers from charging edge providers for using their service, thus forcing them to sell this service to all who ask at a price of $0, we see no room at all for “individualized bargaining.” (citing Cellco) (emphasis added). It’s all about the presumption!
So where does that leave us?
If you want to go back to the case-by-case approach contained in the 2010 OIO—with a presumption against any priority deals—the best option is legislation that resembles the FCC’s 2010 rules. The downside of this option is that it involves Congress.
Alternatively, if you can tolerate a presumption in favor of any priority deal, then you can ground your case-by-case approach in Section 706 along the lines of what the D.C. Circuit telegraphed in its Verizon decision. (Some are uncomfortable with the breadth of the FCC’s reinterpretation of Section 706. On this, see the amicus brief filed on Friday by TechFreedom, CEI, ICLE and leading law and economics scholars in the Sixth Circuit’s case, about the FCC’s attempt to preempt state muni broadband laws.)
Of course, there are hybrid approaches that involve burden shifting, which could be lifted from antitrust, but that’s beyond the scope of this blog.
While most economists could live with the more permissive case-by-case regime, it seems the activists on the Left cannot. In hindsight, the compromise in the 2010 OIO was pretty “just and reasonable.”
With the tumble in ISP capex that just happens to coincide with its Open Internet Order, the FCC is in full damage control. Because there are no economists to defend this hot mess, the activists (and analysts who selectively quote CFOs) have rushed to the agency’s rescue.
The latest defense comes from Free Press, an activist group that specializes in cat-themed protests. Their critique is replete with errors, only some of which I address here.
Some mistakes are trivial. Like the claim that I ignored Comcast in reaching an estimated $3.3 billion in capital flight by the major ISPs in the first half of 2015 (compared to the first half of 2014). In case there was any doubt over which firms were included in my sample, I provided a breakdown of the estimate on slide 4 of my AEI presentation.
Some mistakes are pivotal. Free Press made a significant error in reporting Sprint’s capex: Fix that single mistake and Free Press’s capex results are nearly identical to mine. Even with a new sample of “Top ISPs” (and ignoring the Sprint error), Free Press illustrates that Title II was associated with a three percent decline in ISP capex in the first half of 2015 (compared to the first half of 2014). Is that a good thing? I’ll come back to this one in a bit.
Among other straw man arguments, Free Press asserts falsely that I believe 100 percent of the observed decline in ISP capex is attributable to Title II. The language of my blog post made clear that the relevant and unknown percent was likely positive, and only needed to be three percent (of the observed decline) to swamp the purported benefits of the FCC’s Open Internet Order.
Although I can’t say precisely what portion of the capex decline is attributable to Title II, I can say that it is roughly one-third of the predicted effects of mandatory unbundling on ISP capex (slide 3). This suggests that ISPs are discounting the likelihood that the Open Internet Order will lead to policies mandating unbundling of their residential Internet connections. (With no discounting, the capex decline would be even bigger.) In the face of this murky abyss of regulatory uncertainty, the ISPs appear to be hedging their bets against a worst-case scenario.
Equally silly is Free Press’s claim that the capex declines “were all announced well ahead of the FCC’s reclassification order.” To substantiate this claim, Free Press cites an AT&T press release from November 2012, which states that AT&T’s capex would be “returning to normal levels in 2015.” Free Press conveniently leaves out the next sentence from the press release: “AT&T expects capital spending to be approximately $22 billion for each of the next three years, then return to pre-Project VIP levels.” Oops. That would include through November 2015, but AT&T’s guidance for most of 2015 has been $18 billion, with a slight revision to $19 billion (not including DirecTV) in August. Even more damning to Free Press’s advocacy is the fact that a “return to normal levels” implies a return to AT&T’s average capex from 2010-12, of $20.5 billion, which is well north of the $18 billion guidance assuming zero growth.
Under what is best described as the “ISP As Naif” theory, Free Press asserts that the harms associated with the Open Internet Order could not manifest until the second quarter of 2015: “If Singer’s theory had any validity, the results from the second quarter of 2015 should stand out, as it was the first full reporting period after the FCC’s vote.” This is absurd.
The Title II writing has been on the wall since President Obama took the reigns from Chairman Wheeler on November 10, 2014. And the ISPs got the message. In reaction to the heightened risk of Title II, on November 12, 2014, AT&T’s CEO Randall Stephenson said: “It’s prudent to pause [our investments]. We want to make sure we have line of sight on this process and where these rules could land, and then re-evaluate.” Brian Fung of the Washington Post reported on November 14, 2014 that “Wheeler now finds himself facing a movement that is quickly gaining steam around the President’s position.” The notion that ISPs could not figure out that Title II would become the official FCC policy until the FCC voted in February 2015 is insulting. And even if the ISPs were as naive as Free Press would have you believe, the vote itself took place well before the end of the first quarter.
Free Press conveniently omits the CEO’s investment “pause” in its narrative, focusing instead on the absence of any mention of Title II by AT&T’s CFO in an earnings call—as if that were a test of anything. And don’t get me started on the qualification of CFOs to assess regulatory impacts. Rather than crediting the spin from corporate executives, economists prefer to assess data when testing hypotheses.
There are countless other mistakes. For example, Free Press includes Windstream in its sample of ISPs potentially impacted by the Open Internet Order, despite the fact that Windstream is largely a CLEC. But the biggest mistake involves a miscalculation of Sprint’s capex change from the first half of 2014 to the first half of 2015. Free Press claims this increase was a staggering $1.7 billion, which if true would partially offset AT&T’s $3.3 billion capex decline.
With a seemingly large offset for Sprint, Free Press estimates that capex for the “Total Top ISPs” fell by only $1 billion in the first half of 2015—see the last row in the Free Press table—compared to my estimated $2.5 billion decline (including Sprint and T-Mobile). It seems silly that we are arguing about the extent of the decline, but by fudging on Sprint, Free Press can claim (falsely) that capex increased in the second quarter of 2015 relative to the second quarter of 2014. Recall that the second quarter is the relevant one per Free Press’s ISP as Naif theory.
The problem with Free Press’s number for Sprint is technical and concerns a change in accounting: In the fourth quarter of 2014, Sprint began capitalizing the cost of its leased devices. By the first half of 2015, there was $983 million of capitalized leased equipment on the books ($439 million for the calendar first quarter plus $544 million for the calendar second quarter).
There were no such capitalized leased devices in the comparable period for 2014. If you back out capitalized devices from 2015 to get an apples-to-apples comparison, Sprint’s network investment increased from $2.734 billion in the first half of 2014 to $3.410 billion in the first half of 2015, or an increase of $676 million. Not too shabby. Yet that’s a far cry from the $1.7 billion increase presented in the Free Press table.
This is not to say that Sprint is doing anything shady. The cable industry has been capitalizing the cost of leased set-top boxes forever. But including these numbers creates a misleading comparison of Sprint’s 2015 capex with its 2014 capex. Correcting this single error in Free Press’s table results in a $2 billion loss in capex across the “Top ISPs” during the first half of 2015 (compared to the first half of 2014), or a decline of six percent. Compare this to my estimated eight percent decline for major broadband ISPs including Sprint and T-Mobile. Free Press is generating an amazing amount of vitriol over two percentage points of difference—both in the negative territory.
Put differently, nearly $1 billion of Sprint’s alleged increase in capex is simply due to a business model shift (and associated accounting change) rather than increased network investment. For the second quarter numbers, which Free Press believes is the true test of Title II under the “ISP As Naif” theory, removing Sprint’s capitalized handsets from the calculation flips their results: Title II is still associated with a decline of $400 million across the “Top ISPs” (equal to $17.1 billion less $17.5 billion), or a decline of two percent in the second quarter of 2015 (compared to the second quarter of 2014).
Are we still bickering about whether the Open Internet Order had a negative impact on ISP investment?