What Financial Data Shows About the Impact of Title II on ISP Investment

This post kicks of a series of blogs examining some of the more pernicious myths and misunderstandings in telecom policy. With a new fire lit under the net neutrality warriors, misinformation runs rampant and spreads quickly. There is plenty in the record that deserves correcting.

At the same time, we will also highlight the best arguments made by those we generally disagree with in the spirit of trying to tamp down ever-escalating rhetoric in these debates. If policymakers are going to make any progress on developing a consensus-driven, stable, long-term framework for broadband regulation, it is going to require charity, cooperation, and willingness to seek common ground—all of which are in short supply these days.

With that, let’s turn to the first myth: that financial data shows that Title II isn’t hurting Internet service providers’ investment in their networks.

We live in a big, diverse society. Stark divides in policy preferences can be expected. But occasionally there are more banal disagreements over simple empirical questions. Of course, this is Washington, DC, where lies, damned lies, and statistics can go a long way, but where there is an actual fact of the matter, we should be able to get to the bottom of the issue.

Enter the most recent back and forth in the net neutrality debate: whether broadband capital investment is up or down after the 2015 Open Internet Order’s Title II designation. Economic impact of any policy decisions is of course a nuanced and complicated inquiry, and any serious analysis would at least attempt to hold for several other variables constant and look only at what investment would have been if the policy had been different.

Many have weighed in on the issue, but the most distinct sides of this debate have been laid out by advocacy group Free Press on one side, and economist Hal Singer on the other. Free Press argues that aggregate broadband investment increased by 5.3 percent in 2015 and 2016 relative to 2013 and 2014, whereas Singer argues there was a 5.6 percent decline relative to 2014 levels. What is odd is that these two are using more or less the same financial data, over the same time period, but come to different results—why does one see up where the other sees down?

First, let’s note this is a pretty superficial question as far as policy analysis goes. Whether investment is up or down after Title II classification doesn’t necessarily tell us much about to what extent Federal Communications Commission (FCC) policy is to thank (or blame) for those changes. Not only is the time period far too short, investment overall could very well be up, but not up as much is it otherwise would be without Title II (same if there was a decline). But these counterpoised, if simplistic, narratives are out there. What’s more, the FCC placed major (but undue) emphasis on this question in the recently released open Internet Notice of Proposed Rulemaking (NPRM), the Internet Association has been relying on Free Press’s figures for their own advocacy, and the media has enjoyed the finger pointing, so it’s worth clarifying the disagreement.

Let’s start with Free Press’s own data for 2014, 2015, and 2016 (line 1):

Then, take two simple steps to make the Free Press’s and Singer’s data sets comparable. First, Singer only looks at the 12 largest ISPs, whereas Free Press examines 24 (and counts Charter and Time Warner Cable together for the period), so we subtract out the ISPs not in Singer’s analysis (line 2i). Removing the 13 smaller ISPs from Free Press’s numbers doesn’t change the overall trend that they find—in fact investment by smaller ISPs is disproportionately down after the Open Internet Order, so normalizing the data actually makes the Free Press narrative stronger. Second, Singer used Sprint’s year-to-date financial data that was available at the time, whereas Free Press used a full year, so we subtract out Sprint’s first quarter investment that was not in Singer’s analysis (line 2ii). Again, this doesn’t change the arc of either’s narrative—we’re simply normalizing the two sets to find where they diverge.

Next—and this is the important part—Singer controlled for three of the most obvious external factors that have nothing to do with Title II that must be subtracted from Free Press’s figures: (1) the mid-period change in how Sprint treats handsets for accounting purposes (line 3i), (2) AT&T’s investment in Mexico (line 3ii), and (3) AT&T’s investment in DirecTV (line 3ii).

Sprint’s decision to capitalize its handsets is a simple change in their accounting, that doesn’t materially affect their business. There is absolutely nothing wrong with capitalizing customer equipment like handsets (though some financial analysts have noted this “change is more of an issue of accounting allocation” that “artificially flattered” Sprint’s margins). Our point here is simply that you must control for this mid-period change as a matter of basic analytical rigor.

Similarly, with DirecTV, it wasn’t counted in the mix in AT&T’s capital expenditures in 2014, but then it was included in 2015 and 2016. These video investments are not really the type of broadband investment we would consider impacted by Title II (if anything, it represents capital flight away from Title II services), but more importantly, this is an artificial change halfway through the time period that should be controlled for. Same goes for AT&T’s investment abroad. The competitive dynamics of Mexico’s telecom industry are generally unaffected by U.S. regulatory decisions.

Once you control for these three simple factors, the differences between Free Press and Singer become much more narrow:

So even if you don’t want to attempt a more sophisticated economic analysis that looks only at changes due to Title II, and instead simply look to whether investment is up or down after the 2015 Open Internet Order and only control for the three-most obvious external factors, financial filings shows broadband investment went down roughly 2-3 percent after the Open Internet Order, consistent with industry’s own findings.

Capital investment in networks is not the be-all and end-all of net neutrality policy. We care about investment throughout the entire ecosystem, as well as the capacity for innovation, and the ability for networks to continue evolving to meet new demands and support new applications. And again, what we should worry about is the effect of Title II itself, not purely whether investment is up or down. Other variables, such as the natural ending of specific investment programs, or cost-savings from new technology, may be a considerable explanation of the 2-3 percent decline (and those who overly emphasize the investment decline in support of removing Title II classification should fear being hoist by their own petard should these trends continue after Pai returns to Title I).

But it is legitimate to be concerned about how particular regulatory structures impact the level of investment. It’s especially important that we see continued investment in the infrastructure that supports “best-efforts” open Internet. And there is good reason to think Title II would affect this. Not only did the Open Internet Order take potential business models off the table, and throw others into uncertainty under the Internet Conduct Standard, it represents the first step down the slippery slope to more onerous utility regulations, such as network unbundling requirements or price regulation.

Of course the FCC forbore from those scarier sections of Title II, and it is unlikely that even a Wheeler FCC would have implemented them. But considering large capital expenditure projects are recovered over a decade-plus time horizon, and the convulsions of populist politics the country is going through, this camel’s nose should not be taken lightly. As investment analyst Craig Moffett put it when he downgraded his rating on cable stocks after the Open Internet Order, “It would be naïve to suggest that the implication of Title II … doesn’t introduce a real risk of price regulation.”

Jason Furman, then-chairman of the White House Council of Economic Advisers, explained in 2013, “investments in infrastructure depend critically on a stable, predictable, and light touch regulatory regime.” While there are many legitimate goals of regulation, he said, this need for stability and predictability has historically been “the motivation for the approach this administration and the Federal Communications Commission have taken in a wide range of areas like the Open Internet.”

Chairman Pai’s NPRM looks to at least return to a light-touch regime, pulling us out of Title II, but I fear this step will be one more swing in the long back-and-forth of net neutrality—we still need that stability and predictability. We should be looking to put down our sticks and stones and work towards legislation.