Articles by Jerry Ellig

Jerry Ellig is a senior research fellow at the Mercatus Center at George Mason University. He has also served as deputy director of the Office of Policy Planning at the Federal Trade Commission and as a senior economist at the Joint Economic Committee of the US Congress.

Federal Communications Commission (FCC) Chairman Ajit Pai today announced plans to expand the role of economic analysis at the FCC in a speech at the Hudson Institute. This is an eminently sensible idea that other regulatory agencies (both independent and executive branch) could learn from.

Pai first made the case that when the FCC listened to its economists in the past, it unlocked billions of dollars of value for consumers. The most prominent example was the switch from hearings to auctions in order to allocate spectrum licenses. He perceptively noted that the biggest effect of auctions was the massive improvement in consumer welfare, not just the more than $100 billion raised for the Treasury. Other examples of the FCC using the best ideas of its economists include:

  • Use of reverse auctions to allocate universal service funds to reduce costs.
  • Incentive auctions that reward broadcasters for transferring licenses to other uses – an idea initially proposed in a 2002 working paper by Evan Kwerel and John Williams at the FCC.
  • The move from rate of return to price cap regulation for long distance carriers.

More recently, Pai argued, the FCC has failed to use economics effectively. He identified four key problems:

  1. Economics is not systematically employed in policy decisions and often employed late in the process. The FCC has no guiding principles for conduct and use of economic analysis.
  2. Economists work in silos. They are divided up among bureaus. Economists should be able to work together on a wide variety of issues, as they do in the Federal Trade Commission’s Bureau of Economics, the Department of Justice Antitrust Division’s economic analysis unit, and the Securities and Exchange Commission’s Division of Economic and Risk Analysis.
  3. Benefit-cost analysis is not conducted well or often, and the FCC does not take Regulatory Flexibility Act analysis (which assesses effects of regulations on small entities) seriously. The FCC should use Office of Management and Budget guidance as its guide to doing good analysis, but OMB’s 2016 draft report on the benefits and costs of federal regulations shows that the FCC has estimated neither benefits nor costs of any of its major regulations issued in the past 10 years. Yet executive orders from multiple administrations demonstrate that “Serious cost-benefit analysis is a bipartisan tradition.”
  4. Poor use of data. The FCC probably collects a lot of data that’s unnecessary, at a paperwork cost of $800 million per year, not including opportunity costs of the private sector. But even useful data are not utilized well. For example, a few years ago the FCC stopped trying to determine whether the wireless market is effectively competitive even though it collects lots of data on the wireless market.

To remedy these problems, Pai announced an initiative to establish an Office of Economics and Data that would house the FCC’s economists and data analysts. An internal working group will be established to collect input within the FCC and from the public. He hopes to have the new office up and running by the end of the year. The purpose of this change is to give economists early input into the rulemaking process, better manage the FCC’s data resources, and conduct strategic research to help find solutions to “the next set of difficult issues.”

Can this initiative significantly improve the quality and use of economic analysis at the FCC?

There’s evidence that independent regulatory agencies are capable of making some decent improvements in their economic analysis when they are sufficiently motivated to do so. For example, the Securities and Exchange Commission’s authorizing statue contains language that requires benefit-cost analysis of regulations when the commission seeks to determine whether they are in the public interest. Between 2005 and 2011, the SEC lost several major court cases due to inadequate economic analysis.

In 2012, the commission’s general counsel and chief economist issued new economic analysis guidance that pledged to assess regulations according to the principal criteria identified in executive orders, guidance from the Office of Management and Budget, and independent research. In a recent study, I found that the economic analysis accompanying a sample of major SEC regulations issued after this guidance was measurably better than the analysis accompanying regulations issued prior to the new guidance. The SEC improved on all five aspects of economic analysis it identified as critical: assessment of the need for the regulation, assessment of the baseline outcomes that will likely occur in the absence of new regulation, identification of alternatives, and assessment of the benefits and costs of alternatives.

Unlike the SEC, the FCC faces no statutory benefit-cost analysis requirement for its regulations. Unlike the executive branch agencies, the FCC is under no executive order requiring economic analysis of regulations. Unlike the Federal Trade Commission in the early 1980s, the FCC faces little congressional pressure for abolition.

But Congress is considering legislation that would require all regulatory agencies to conduct economic analysis of major regulations and subject that analysis to limited judicial review. Proponents of executive branch regulatory review have always contended that the president has legal authority to extend the executive orders on regulatory impact analysis to cover independent agencies, and perhaps President Trump is audacious enough to try this. Thus, it appears Chairman Pai is trying to get the FCC out ahead of the curve.

The National Transportation Safety Board recommended yesterday that states ban all non-emergency use of portable electronic devices while driving, except for devices that assist the driver in driving (such as GPS). The recommendation followed the NTSB’s investigation of a tragic accident in Missouri triggered by a driver who was texting.

Personally I don’t see how someone can pay attention to the road while texting. (I’m having a hard enough time paying attention to a conference presentation while I’m typing this!) But the National Transportation Safety Board’s recommendation is a classic example of regulatory overreach based on anecdote.  The NTSB wants to use one tired driver’s indefensible and extreme texting (which led to horrific results) as an excuse to ban all use of portable electronic devices while driving – including hands-free phone conversations.  Before states act on this recommendation, they should carefully examine systematic evidence – not just anecdotes — to determine whether different uses of handheld devices pose different risks. They should also consider whether bans on some uses would expose drivers to risks greater than the risk the ban would prevent.

The Senate might vote this week on Sen. Hutchison’s resolution of disapproval for the FCC’s net neutrality rules.  If ever there was a regulation that showed why independent regulatory agencies ought to be required to conduct solid regulatory analysis before writing a regulation, net neutrality is it.

For more than three decades, executive orders have required executive branch agencies to prepare a Regulatory Impact Analysis accompanying major regulations.  One of the first things the agency is supposed to do is identify the market failure, government failure, or other systemic problem the regulation is supposed to solve. The agency ought to demonstrate a problem actually exists to show that a regulation is actually necessary.

But the net neutrality rules have virtually no analysis of a systemic problem that actually exists, and no data demonstrating that the problem is real.  Instead, the FCC’s order outlines the incentives Internet providers might face to treat some traffic differently from other traffic, in a discussion heavily freighted with “could’s” and “may’s”.  Then it offers up just four familiar anecdotes that have been used repeatedly to support the claim that non-neutrality is a significant threat  (all four fit in paragraph 35 of the order).  The FCC asserts without support that Internet providers have incentives to do these things even if they lack market power, and indeed in a footnote it dispenses with the need to consider market power: “Because broadband providers have the ability to act as gatekeepers even in the absence of market power with respect to end users, we need not conduct a market power analysis.” (footnote 87)

Thus far, no administration of either party has sought to apply Regulatory Impact Analysis requirements to independent agencies. If administrations won’t, Congress should.


Federal Communications Chairman Genachowski previewed the universal service reform plan the commissioners are discussing in a speech today.

The speech offers a masterful summary of the myriad inefficiencies created by the current universal service subsidies and intercarrier compensation payments. Most of the examples highlight plain old-fashioned waste. The universal service program collects billions of dollars from telephone subscribers, then simply wastes a goodly portion of it by subsidizing telephone competition in places where unsubsidized service from cable or satellite already exists, subsidizing multiple mobile wireless competitors, and subsidizing local phone companies that have little incentive for cost containment because they are still subject to rate-of-return regulation. The intercarrier compensation system uses per-minute charges to collect billions of dollars from telephone subscribers and hands it to phone companies that sometimes charge as little as $8 a month for phone service. There’s also a race to game this system as the companies that benefit seek new ways to inflate the regulated charges they collect, and the companies that pay seek clever ways to avoid paying.

It’s a powerful brief for reform. Never thought I’d live to see the day whan an FCC chairman would say so many things that are substantiated by economic research.

Nevertheless, a few parts of the speech give me cause for concern about the solutions the FCC commissioners may be discussing.

First, the chairman claims that 18 million Americans live in areas without access to broadband — up from the 14 million estimated in the National Broadband Plan.  The size of this figure suggests to me that the FCC is still over-estimating the number of people without access by defining “broadband” as a speed fast enough to exclude 3G wireless, many small rural Wireless Internet Service Providers, and satellite. Absent an adjustment in the definition of broadband, the subsidy program will be larger than it needs to be, and so telephone consumers will pay excessive universal service charges. Continue reading →

I can’t help but think that there might be  a big advantage of having the AT&T-T-Mobile merger go to court.  For once, the high-profile action everyone pays attention to will occur in an antitrust forum where the decision criterion is the effects of the merger on consumer welfare, period.   Regardless of what one thinks about the merger, it’s nice to see that we’ll finally have a knock-down, drag-out fight based on whether a big telecommunications merger harms consumers and competition.  That’s the antitrust standard the Department of Justice has to satisfy in order to prevent the merger. 

This will be a refreshing change from the Federal Communications Commission’s “public interest” standard, which allows the commission to object on grounds other than consumer welfare and demand all manner of concessions that have nothing to do with remedying anticompetitive effects of a deal. Case in point: Comcast must now offer broadband service for $9.95 per month to low-income households as a condition for getting approval to buy 51 percent of NBCUniversal. Now, I’m all for seeing low-income households get access to broadband, but subsidizing one subset of customers has little to do with mitigating any possible anticompetitive effects of allowing a cable company to own NBCUniversal. As FCC Commissioners McDowell and Baker said in their statement on that transaction, “Any proposed remedies should be narrow and transaction specific, tailored to address particular anti-competitive harms. License transfer approvals should not serve as vehicles to extract from petitioners far-reaching and non-merger specific policy concessions that are best left to broader rulemaking or legislative processes.” 

In short, if AT&T wins in court, the FCC should approve the merger promptly without additional conditions.

A couple days before Congress announced a debt deal, half a dozen telecommunications companies filed a plan on July 29 with the Federal Communications Commission that attempts to resolve a much longer-running set of negotiations over big bucks.  The “America’s Broadband Connectivity” Plan seeks to replace Universal Service Fund subsidies for telephone service in rural areas with subsidies for broadband in rural areas.

Like the federal budget negotiations, the never-ending negotiations over USF get bogged down in arguments over distribution: who gets what.  Indeed, it’s almost exclusively an argument over which companies get what. But federal telecommunications policy is supposed to advance the overall public interest, not just haggle over what corporate interest gets what piece of my pie. Here is a quick take on the biggest strengths and weaknesses of the plan in terms of advancing overall consumer welfare. By “consumer welfare,” I mean not just the welfare of the folks receiving subsidized services, but also the welfare of the majority who are paying a 15 percent charge on interstate phone services to fund the USF.


Fixed-term commitment: Rural phone subsidies have become a perpetual entitlement with no definition of when the subsidies can end because the problem is considered solved.  The ABC plan proposes a 10-year commitment to rural broadband subsidies.  By 2022 the FCC should assess whether any further high-cost universal service program is needed. This idea remedies a significant deficiency in the current high-cost subsidy program, which doesn’t even have outcome goals or measures. (That’s why I like to sing the final verse from “And the Money Kept Rolling In” from Evita when I talk about universal service.  Free State Foundation President Randy May asked me for an encore of this at the end of the foundation’s July 13 program on universal service, available here.)

Intercarrier compensation: “Intercarrier compensation” refers to the per-minute charges communications companies pay when they hand off phone traffic to each other. The plan proposes to ramp down all intercarrier charges to a uniform rate of $0.0007/minute.  Economists who study telecommunications have pointed out for decades that high per-minute charges reduce consumer welfare by discouraging consumers from communicating as much as they otherwise would.  MIT economist Jerry Hausman, in a paper prepared for the filing, estimates that low, uniform intercarrier charges would increase consumer welfare by about  $9 billion annually.

Legacy obligations: Public utility regulation traditionally forced regulated companies to offer certain services or serve certain markets at a loss, then charge profitable customers higher prices to cover the losses. Judge Richard Posner referred to this opaque practice as “Taxation by Regulation“: the customers paying inflated prices get “taxed” to accomplish a public purpose, but they don’t know it.  Some of these obligations continue today as federal requirements applied to “Eligible Telecommunications Carriers” or state “Carrier of Last Resort” obligations.  The plan would remove these obligations for companies that are not receiving USF subsidies.


Definition of broadband: The plan would continue to inflate the cost of rural broadband subsidies by defining “broadband” as 4 megabytes per second download and 768 kilobytes per second upload.  This means 3G wireless, satellite, and some wireless Internet service providers do not count as “broadband.” This decision more than doubles the number of households considered “unserved” and rules out some lower-cost technologies.  Jerry Brito and I have written extensively about both the economics and the legality of this.  Interestingly, the ABC coalition’s legal white paper arguing that the commission has legal authority to adopt the plan makes no effort to show that the commission has authority to subsidize 4 mbps broadband; it only shows the commission has authority to subsidize some form of broadband.

Alternative cost technology threshold: The plan includes an “alternative cost technology” threshold that allows substitution of satellite broadband for customers who would cost more than $256/month to serve.  Inclusion of a threshold is actually a strength. But the $256/month figure is way too high.  Satellite broadband with speeds of 1-2 mbps is now available for $60 – $110 per month.  Consumers who pay a 15 percent surcharge on their local phone bills to fund USF should not be expected to provide a subsidy of more than $200 per month.

Mobility: The plan appears to advocate subsidies for mobile broadband service in places where it is not currently available.  So now the rural entitlement expands to include not just basic broadband service in the home to stay connected, but also a mobile service that a lot of Americans don’t even buy unless their employers pay for it! I question whether mobile broadband satisfies the 1996 Telecommunications Act’s criteria for universal service subsidies, such as “essential” (not just nice) for education or public safety, or subscribed to by a “substantial majority” of households. These questions should be thoroughly examined before anyone receives subsidies for mobile broadband. At a minimum, households should be eligible for only one broadband subsidy — wireline or mobile — but not both.



While most folks have been obsessing over their income taxes the past few weeks, Jerry Brito and I have been obsessing about a non-tax: the universal service assessments on our phone bills.

More specifically, the Federal Communications Commission has asked for comments on its plan to gradually turn the current phone subsidy program in high-cost rural areas into a broadband subsidy program in high-cost rural areas. This opens up a big tangled can of worms.  Comments are due Monday.  We deal with two issues in our comment:

Definition of broadband: Thankfully, the FCC is asking for comments on its proposal to define broadband as 4 Mbps download/1 Mbps upload. This is an important decision with a big effect on the size of the program. The 4 Mbps definition more than doubles the number of households considered “unserved,” because it doesn’t count 3G wireless or slower DSL or slower satellite broadband as broadband. It also raises the cost of the subsidies by requiring more expensive forms of broadband.

The definition fails to fit the factors the 1996 Telecom Act says the FCC is supposed to consider when determining what communications services qualify for universal service subsidies.  A download speed of 4 Mbps is not “essential” for online education; most online education providers say any broadband speed or even dialup is satisfactory. Nor is that speed “essential” for public safety; the biggest barrier to public safety broadband deployment is creation of an interoperable public safety network, which has nothing to do with USF subsidies. And the proposed speed is not subscribed to by a “substantial majority” of US households.  The most recent FCC statistics indicate that the fastest broadband download speed subscribed to by a “substantial majority” of US households is probably 768 kbps.

Definition of performance measures: Fifteen years after passage of the legislation that authorized the high cost universal service subsidies, the FCC has proposed to measure the program’s outcomes.  Actually, the FCC wants to measure intermediate outcomes like deployment, subscribership, and urban-rural rate comparability — not ultimate outcomes like expanded economic and social opportunities for people in rural areas.  But it’s a start …  provided that the FCC actually figures out how the subsidies have affected these intermediate outcomes, rather than just measuring trends and claiming the universal service subsidies caused any positive trends observed.  We have some suggestions on how to do this.  

Our full comment is available here.

In previous posts, I’ve criticized the Federal Communications Commission for arbitrarily jacking up the speed in its definition of broadband (to 4 mbps download/1mbps upload) so that third generation wireless does not count as broadband. This makes broadband markets appear less competitive.  It also expands the “need” for universal service subsidies for broadband, since places that have 3G wireless but not wired broadband get counted as not having broadband.

The FCC’s definition is based on the speed necessary to support streaming video.  I rarely watch video on my computer. But tonight I had a chance to test the wisdom of the FCC’s definition.  I’m in rural southern Delaware with broadband access only via a 3G modem. I wanted to watch more State of the Union coverage than the broadcast channels out here carried. So, I fired up the old PC and watched things on  The video showed up fine and smooth, and it didn’t even burp when I opened another window to start working on this post.

So now I have not just analysis that questions the FCC’s definition of broadband, but that most precious of commodities in Washington regulatory debates: AN ANECDOTE!!!

The House and Senate have now both passed bills aimed at encouraging telework in the federal government. As anyone who has had to commute to work in the Washington DC area knows, the national capital area could probably use a good dose of telework to relieve traffic congestion.

According to Joe Davidson’s column in the Washington Post, “The inability or unwillingness of supervisors to manage staff members they can’t see has long been cited as a major reason” more federal employees don’t telework. This fits with what I’ve heard from some current or former federal managers.  “I have enough trouble getting work out of people when they’re in the office,” one remarked.

The legislation offers some simple solutions: Tell federal agencies they have to allow all employees to work remotely unless there’s some reason a position isn’t conductive to telework. And accompany that with training so that managers will be better equipped to manage employees who aren’t in the office.

I’m a big fan of telework. But one of the keys to making it work is holding employees accountable for results instead of inputs like time on task or time hanging around the office.  It’s possible to do this even when the desired results are hard to measure.  Universities, for example, evaluate professors based on the quality of their teaching and research, not the number of hours they spend preparing for class or writing. This system is hardly perfect, and some places do this better than others. But on balance, it works much better than telling professors they’ve fulfilled their obligation by showing up at the office 40 hours a week.

So the key question in making telework work in the federal government is, “How well do agencies hold individual employees accountable for results?”  Here, the federal government has a few handicaps to overcome. It’s hard to fire people for poor performance.  Pay is set by the federal pay scale, which does not necessarily create a direct link between pay and the value of the employee’s accomplishments to taxpayers. And agencies do not always create a clear understanding of how the individual employee’s contribution affects the results the agency is supposed to produce.

Granted, the federal government is probably better at dealing with some of these challenges now than it was 20 years ago, especially for the senior executive service. But most federal jobs are still a long way away from at-will employment with clear performance measures tied to the organization’s goals. This is a change that requires not just “more training” or “cultural transformation,” but also a redefinition of the terms of federal employment.

Given those circumstances, I think federal managers are justified in their concern that giving most employees the automatic right to telework could reduce productivity.  I can think of two ways to make telework work in the current federal employment environment:

1. Make people earn it. Employees who show they can get things done without a lot of supervision in the office are the most obvious candidates to be effective working remotely.

2. Mandate a trial period and evaluation. If you think it’s fair to guarantee the opportunity to telework to most employees, mandate only that it must be offered on a trial basis. Continuation depends on performance.

These are, of course, second-best solutions.  And there may be others.

In a 3-2 vote, the Federal Communications Commission recently decided to jack up its official definition of “broadband” from 200 kbps download to the 4 mbps dpwnload/1 mbps upload used as a benchmark in Our Big Fat National Broadband Plan. The three commissioners in the majority also declared that the definition of broadband will continue to evolve as consumers purchase faster connections to utilize new applications.

Several months earlier, the FCC launched a proceeding to figure out how to convert universal service subsidies for rural telephone service into universal service subsidies for rural broadband service.  Put these two decisions together, and it looks like the majority on the FCC is hell-bent on establishing rural broadband subsidies as a perpetual entitlement program that will never “solve” the rural availability problem because the goalposts will keep moving.

The current USF program taxes price-sensitive services (long distance and wireless) to subsidize a service that is not very price sensitive (local phone connections).  If the FCC takes a further step on the funding side and starts collecting universal service assessments from broadband, it will diminish broadband subscribership by taxing a service that is even more price sensitive: broadband connections. (I explained this a few months ago here.)

It’s time to get off this merry-go-round. The solution was suggested by MIT economist Jerry Hausman back when the FCC first started creating the current universal service programs in response to the Telecom Act of 1996: use revenues from spectrum auctions. 

Instead of having the FCC perpetually collect assessments from broadband or telephone services to subsidize broadband buildout in rural areas, Congress should earmark revenues from the next spectrum auction for one-time buildout grants in high-cost areas. The grants should be awarded via a competitive procurement auction that would force subsidy-seekers in different locations to compete with each other for the federal dollars. And Congress should explicitly wind down the universal service telephone subsidies in high cost areas and prohibit the FCC from using universal service assessments to fund broadband deployment in these places.

Using revenues from spectrum auctions would avoid the distortions and perverse consequences caused by ongoing universal service assessments on broadband or telephone services. One-shot deployment grants would ensure that the availability problem gets solved, so the federal government can declare victory and get out of the perpetual subsidy business.

Of course, some locations in the US are so expensive to serve that the potential revenues might not even cover the operating costs of broadband. But it does not follow that operators in these places need an ongoing stream of subsidies. When preparing their subsidy bids, they will have to calculate how large the one-shot payment needs to be to induce them to take on the capital costs and the ongoing operating costs. In other words, they can bank some of the one-shot subsidy and use it to cover the difference between revenues and operating costs.

This modest proposal does not address all aspects of the universal service fund. But it would achieve a clear objective — bringing broadband to rural areas — while allowing the FCC to extricate itself from the business of distributing $4.6 billion a year in subsidies. Let’s see a timetable for withdrawal!