Implications of Mozilla for Agency Economic Analysis, by Jerry Ellig
The D.C. Circuit released its opinion in Mozilla v. FCC, which largely upheld the Federal Communications Commission’s (FCC’s) Restoring Internet Freedom order, on October 1, 2019. Much of the ensuing commentary and policy debate will no doubt center on the fact that the court upheld the substance of the order but struck down the provision that pre-empted states from imposing their own net neutrality rules.
But the court also devoted 27 pages of its 146-page decision to dissecting the FCC’s economic analysis, to see if the petitioners raised any challenges that would trip the “arbitrary and capricious” wire. No need to ask “Where are the Economists?,” as Aaron Nielson did in his response to the DC Circuit’s 2016 net neutrality decision.
This part of the court’s opinion was game, set, and match for the FCC. And it holds important lessons for agencies in how they conduct and present the economic analysis that informs regulatory decisions.
Title II classification and investment
A major issue in the net neutrality debate is whether classifying broadband as a common carrier under Title II of the Communications Act of 1934 affects investment incentives for broadband Internet service providers (ISPs). The court concluded that the commission’s finding that Title II classification likely reduced broadband investment was supported by substantial evidence (p. 74). The court declared that “Predictions regarding the actions of regulated entities are precisely the type of policy judgments that courts routinely and quite correctly leave to administrative agencies (p. 74). The opinion then proceeded to scrutinize the economic studies on which the FCC relied in some detail.
The court said it was appropriate for the FCC to rely on a study by Hal Singer that used annual industry data to show that major ISPs’ investment fell by 5.6 percent between 2014 and 2016. Mozilla and intervenors pointed to a Free Press study by Derek Turner using data from publicly-traded ISPs that found investment accelerated after the 2015 Title II order. The court noted that the FCC also considered the Free Press study, plus a third study by Doug Brake that adjusted both the Singer and Free Press data to use the same time period, cover the same ISPs, and remove investments unaffected by the 2015 regulatory change. This third study found that both the Singer and the Free Press data showed investment declined by 2-3 percent per year in 2015 and 2016. Given this convergence of results, plus the fact that the commission assigned “quite modest probative value” to studies that simply compared annual investment figures, the court found the FCC’s reliance on this study to be reasonable. Similarly, the petitioners attacked a Free State Foundation study that found investment by the 16 largest ISPs deviated from its prior trend starting in 2015, but the court noted that the FCC regarded this study merely as suggestive and confirmatory of other studies (pp. 75-78).
The opinion also discussed an event study by George Ford that found the FCC’s surprise announcement in 2010 that it was considering Title II reclassification was associated with a $30-40 billion annual decline in investment by the broadcasting and telecommunications sector between 2011 and 2015. The court approvingly noted that the FCC made the appropriately modest claim that the Ford study only indicated the direction of investment, not the magnitude, since it covered a broader sector of the economy than just broadband ISPs (pp. 78-79). The opinion also noted criticism of the Ford study and a competing study by Christopher Hooton, but concluded that “The Ford-Hooton dispute seems far to sophisticated for us to credibly take sides” (p. 79) and the criticism is “insufficient ground for us to call the Commission’s finding unreasonable” (p. 80).
The opinion stated in multiple places (pp. 76-77, 81) that the FCC drew extensively on a study by Thomas Hazlett and Joshua Wright, which was based on an econometric “natural experiment” method to assess the effects of past FCC policy changes on DSL subscribership. DSL subscribership shifted upward significantly after the FCC removed DSL line-sharing rules in 2003 and then again when the FCC reclassified DSL as an information service in 2005. The court did not explicitly evaluate this study but stated that it was “uncontroverted by Petitioners” (p. 76).
The court addressed a study by Robert Crandall that found the Title II order was not correlated with a change in ISP stock prices. Petitioners argued that the FCC arbitrarily gave this study little weight. The court responded that the FCC did not behave arbitrarily because the absence of a stock price impact on ISPs need not imply that the Title II order had no effect on ISP investment (pp. 83-84).
On the issue of investment by content producers and other “edge providers,” the court said that it was not arbitrary for the FCC to conclude that the Title II order did not increase investment by edge providers, because studies in the record showed neither a correlation nor a causal relationship between Title II classification of broadband and edge provider investment (pp. 81-82).
Harms to edge providers and consumers
The commission determined that potential harms to edge providers and consumers could be prevented by a combination of disclosure requirements, competition, and antitrust and consumer protection law. The court declared that this judgment was not arbitrary (pp. 85-86).
The court said the fact that very few instances of harm have occurred since the commission first adopted a transparency rule in 2010 is a reasonable justification for the FCC’s transparency rules (p. 87).
The court found the FCC offered valid reasons for believing that there is significant competition in the broadband market. Data in the FCC order show that many broadband customers have access to at least two wireline competitors. Empirical research (cited in the footnotes to paragraph 126 in the Restoring Internet Freedom order) finds that even two competitors significantly constrain each other. In addition, when a broadband ISP operates both in competitive and non-competitive markets, there are reputational and business reasons for it to behave similarly in both types of markets (pp. 88-91). Judge Williams’ separate opinion also notes that the Herfindahl-Hirschmann Index for ISPs offering the fastest broadband speeds – greater than or equal to 25 Mbps download and 3 Mbps download – meets the Department of Justice’s designation of “moderately concentrated” (Williams opinion, p. 23). A “moderately concentrated” industry is clearly not a monopoly.
Although the court said the FCC’s conclusion about competition is reasonable, it also said the commission “fails to provide a fully satisfying analysis of the competitive constraints faced by broadband providers” because it only mentioned, but did not significantly analyze, the competitive constraints imposed by satellite, fixed wireless, and broadband offered at speeds lower than 25 Mbps download and 3 Mbps upload (p. 89).
The court rejected petitioners’ claim that the FCC, without giving a reason, abandoned its previous finding that consumers are victims of a “terminating access monopoly” once they choose a broadband provider. In reality, the FCC stated that many subscribers now use both fixed and mobile connections to access the Internet, so there is no terminating access monopoly for those subscribers. Even where a terminating access monopoly exists, it is not clear whether the resulting prices are economically inefficient (p. 90). (The two-sided nature of the market gives ISPs an incentive to compete more vigorously for subscribers if they can charge content providers monopoly prices for access to the subscribers.)
The court also rejected the claim that the FCC ignored its previous findings that high switching costs deter customers from changing broadband providers. The FCC indicated that customer turnover (the “churn” rate) may not be as low as previously estimated, and in any case a low churn rate may be a sign of successful competitive efforts to retain customers (pp. 89-91).
Antitrust and consumer protection law
Petitioners contended that in relying on antitrust and consumer protection law, the FCC was impermissibly delegating its responsibilities to other agencies. The court disagreed:
But the Commission has not “mechanically accept[ed] the standards” of other laws as satisfying its own. Instead, it has conducted an independent assessment of the degree of problematic conduct that has been and will be committed by broadband providers and whether, as a policy matter, the benefits of restricting that conduct outweigh the costs. A reasonable piece of that policy-making puzzle, then, is an assessment of other regulatory regimes that might already limit the conduct in question. Therefore, it was not impermissible for the Commission to recognize that the Department of Justice and Federal Trade Commission have the ability to police blocking and throttling practices ex post. (p. 92)
Nevertheless, the court said that the FCC provided an “anemic analysis” of antitrust and consumer protection law that only “theorized” why they are superior to Title II regulation, and hence the FCC’s analysis “barely survives arbitrary and capricious review” (p. 93).
The Restoring Internet Freedom order’s economic analysis, including assessment of the benefits and costs of alternatives, is incorporated at the places in the document that discuss each regulatory change. In addition, a separate brief section titled Cost-Benefit Analysis summarizes benefits and costs. Petitioners offered two types of objections to this analysis. First, they contended that the Notice of Proposed Rulemaking (NPRM) committed the commission to a quantitative benefit-cost analysis because the NPRM said the commission proposed to follow the guidelines in OMB Circular A-4, but the discussion in the benefit-cost section of the order is qualitative. The court pointed out that “…[T]he NPRM made clear that the Commission was not wedded to the idea of following the Circular, and the Circular itself calls for a qualitative analysis under circumstances that the Commission reasonably invoked.” (p. 114)
Petitioners also argued that the benefit-cost analysis did not account for the cost to consumers of reduced innovation and democratic discourse that they claim the Restoring Internet Freedom order would cause. On the contrary, the court pointed out, the FCC showed that the number of instances of potential harm was small; competition protected by antitrust tends to promote free expression as long as customers value free expression; and the transparency rule, ISP commitments, and FTC consumer protection enforcement are intended to preserve free expression (pp. 116-17). The beneficial outcomes the FCC foresees from the order, such as increased broadband investment, expanded service to unserved areas, and greater freedom to offer innovative services, are likely to increase innovation and democratic participation (pp. 118-119).
Lessons for economic analysis
This decision offers at least four lessons for economists at regulatory agencies to consider when they conduct economic analysis and for attorneys and other authors of orders and regulatory preambles to consider when they write the agency’s justifications for regulations:
- Rely on high-quality (preferably peer-reviewed) analysis. A significant FCC advantage was its reliance on the study by Thomas Hazlett and Josh Wright, which was published in the Review of Industrial Organization, a peer-reviewed economics journal. The results of this study were not even contested, and the court referred to it several times. Another study on which the commission relied, by George Ford, was later published in Applied Economics, also a peer-reviewed journal. FCC economists searched for articles in peer-reviewed economics journals that assessed the effects on economic welfare of blocking, throttling, and paid prioritization – the three practices banned by the Title II order. Apparently this research was so strong that it was not even challenged in court. (But for a discussion of this research, you can go here and here.)
- Honesty is the best policy. On multiple occasions the court approvingly noted how the FCC forthrightly acknowledged the limitations of some of the studies it relied on and was careful not make excessive claims about what they proved. For example, the court stated that “the Commission was clearly assigning quite modest probative value to studies attempting to draw links between the Title II Order and broadband investment…” (p. 78), and the commission was praised for “focusing on what is ‘likely’ to happen, repeatedly flagging shortcomings in the studies it cites, and qualifying their probative force” (p. 84).
- Translate economics (and statistics) into plain English. The court declined to express an opinion on the relative merits of studies by George Ford and Christopher Hooton, deeming the issue too complex for judges to sort out. The issue was actually not that complicated; the question is whether it is legitimate to use forecast data for critical years as an input in a statistical analysis that assesses whether a policy change is correlated with outcomes. But it would be difficult for a layman to render a judgment about who is right based on the information contained in the Restoring Internet Freedom order. Mea culpa for not insisting on a more careful explanation when the order was drafted; I know the FCC’s economists explained this issue to me clearly!
- Showcase your work. I take exception to the court’s conclusion that the FCC’s analysis of antitrust and consumer protection was “anemic,” because I know what is in the FCC order and in the record on this topic. The FCC order discusses the FTC’s prosecution of TracFone for throttling (¶ 141). The order also notes that ISPs’ public commitments not to block or throttle are enforceable under Section 5 of the FTC Act (¶ 142), points to an FTC report which suggested that an ISP could be prosecuted for treating traffic from different edge providers differently if it did not inform customer and obtain their consent (¶ 141), and cites empirical economics literature that demonstrates the deterrent effect of antitrust laws (¶ 152). The antitrust and consumer protection analysis also draws heavily on comments submitted by the FTC’s acting chairman, Maureen Ohlhausen, as well as a separate comment from the FTC staff. I can only conclude that no one brought these passages in the order to the court’s attention.
When the Restoring Internet Freedom order was first made public, I wrote that it contained more thorough economic analysis than the prior FCC order on this topic. The court had no reason to opine on whether the economic analysis was better than what the FCC had previously offered. But all of the economics in the order was clearly sufficient to satisfy “arbitrary and capricious” review – in part because the order took care not to overstate what the economic analysis showed.
Jerry Ellig is a research professor at the George Washington University Regulatory Studies Center. He served as chief economist at the FCC from July 2017 thru July 2018 and oversaw the economic analysis in the FCC’s Restoring Internet Freedom order.