The views expressed below are my own, and are in no way associated with the Federal Communications Commission. I did not contribute any work to this proceeding.
Tomorrow morning, the Federal Communications Commission will vote to roll back a commission rule that reclassified Broadband Providers under Title II of the Telecommunications Act to protect Net Neutrality. Order
Aside from the argument that dubious argument that the Commission need not worry about Net Neutrality because there are few documented examples so far, the Order leans heavily on the argument that Title II reclassification reduced broadband investment in the industry.
I’m not going to get into the merits of Net Neutrality or Title II classification here (short answer – I think it’s good), and as the rules will certainly pass tomorrow, it’s clear that I’m not trying to influence the proceeding. But I looked into the analyses cited by the order, and I noticed a common thread.
The $3 billion Assumption:
I noticed a common thread in almost every study cited by the Commission to justify the case that Title II reclassification had scared broadband providers into reducing their broadband investment – the seemed to rely almost entirely on one assumption (two of the studies cited, Kovacs and Ford, don’t rely on this assumption, I’ll discuss those at the bottom).
Aside from Ford and Kovacs, EVERY analysis mentioned in the order that alleges a decrease in wireline investment makes this assumption or cites to an analysis that makes this assumption.
On face, it might make sense. AT&T acquired DirecTV midway through 2015. Their investment figures pre-2015 did not include DirecTV’s investment, and their investments post 2015 did include DirecTV’s investment, so sensibly that investment needs to be deducted from the later time periods. The problem is that all of these analyses assume that because DirecTV’s capital investment was roughly $3 billion in years prior to the merger, it must have necessarily continued at that level.
BUT WE DONT KNOW. Most of the studies conclude that the industry drop in investment since 2014 was in the area of $3-$4 billion, so this assumption makes up a huge portion of the eventual findings. It’s entirely possible that after the acquisition, AT&T decided to scale back investment in DirecTV’s satellite operations, but we have no way of knowing.
Capital Investment is a Signal
For their part, AT&T was happy to announce to shareholders in their annual filings that they’ve increased their investment in infrastructure every year since 2014. By not disclosing their capital investments in their DirecTV operations, AT&T has found a way to simultaneously tell the FCC that the sky is falling while signaling to their shareholders that business is booming.
This is the issue when a few enormous companies dominate industry trends. We’re left to infer investment outlook for a multi-billion dollar industry based on strategic decisions of a few men in a couple boardrooms who are fully aware that their decisions will influence regulators.
FWIW, AT&T themselves indicated in pre-Title II annual filings that they planned to reduce marginal broadband investment after they finished a huge broadband buildout. Still, it’s entirely possible that AT&T and other industry players actually did reduce their broadband investment because they were afraid of regulatory burden. But that’s not certain, and in light of the outpouring of concern from content creators and edge providers across the web, as well as the unheard concerns of innovators who have yet to innovate, a decision this big shouldn’t hinge on an assumption.
Cited studies that didn’t rely on the DirecTV assumption:
Dr. Anna-Marie Kovacs’s analysis argues that wireless investment in broadband has dropped precipitously in response to the Title II regulation. She doesn’t provide a detailed methodology or raw data, so its hard to audit what she did, but it seems above board. I would add however, that while these companies may not have been investing in hard infrastructure, they invested heavily in spectrum during this time period, and I can’t imagine why they would do that if their future revenue outlook was so poor.
Dr. George S. Ford’s analysis was very different from the others cited in the Order, as it uses post-2010 as the treatment period because he asserts that after the 2010 Order, the industry knew that Title II was in play. He uses a diff-in-diff model which mimics an experiment by creating an artificial control group. Basically, he argues that rather than comparing broadband investment year over year, we should compare it to an estimate of what it would have been if the rules hadn’t been implemented. To do this, he constructed an artificial control group by choosing industries unrelated to telecom that closely followed broadband investment trends prior to 2010, and compares these industries post-2010 to actual broadband investment post-2010.
It’s a pretty cool experimental design, but it’s worth looking at the industries that he chose. His four “artificial control group” industries were:
- Machinery Manufacturing
- Computer and Electronic Product Manufacturing
- Plastic and Rubber Product Manufacturing
- Transportation and warehousing
Choices 2 and 4 jump out at me. I’m completely guessing here, and I don’t know how the BEA calculates their industry estimates for these so I may be completely off… but for #2 – this time period saw smartphones grow from niche high-end accessories to near ubiquity, as well as the explosion of electronic accessories surrounding them. For #4, this time-period also coincided with Amazon’s rise to prominence, which I’m sure has had a pretty significant impact on the transportation and warehousing industry.
Not surprisingly, controls 2 and 4 also happened to massively outperform controls 1 and 3, seem to be the driving factor behind his conclusions. His findings show that 2 and 4 both grew by more than 50%, meaning his conclusion that telecom investment decreased is just an argument that telecom didn’t experience the same gigantic surge in investment that 2 other specific industries did.
To Ford’s credit, his analysis showed that telecom underperformed all four of the control group industries, not just 2 and 4. Also, he includes several robustness checks that I have not yet reviewed closely (but plan to). But if I’m right about 2 and 4, his comparison to the remaining control groups, though still statistically significant, doesn’t seem to hold strong economic significance.