Motley Fool’s Daniel B. Kline reported yesterday that, according to a new study by the Consumer Federation of America (CFA), “Four huge companies that control 70% of the digital communications sector have been using their market power to overcharge the typical household about $45 a month.” Alarming if true.
In yet another display of #TechReporting, Motley Fool issued what amounts to a press release for a public interest group clamoring for price regulation. All of CFA’s claims are taken at face value, and the reporter couldn’t be bothered to interview an economist for independent assessment of CFA’s methodology.
Here’s what Kline would have learned had he reached out to this economist: CFA’s finding that Internet service providers (ISPs) and wireless operators overcharge customers lacks economic rigor and is without merit.
Kline reports that for broadband and video bundles, evil ISPs overcharge customers to the tune of $25 per household per month. AT&T and Verizon reportedly overcharge wireless customer $10 per telephone line. At two lines per household, customers are overpaying by $45 per month (equal to the alleged $25 overcharge on broadband and video plus 2 x the alleged $10 overcharge on wireless telephony).
What is the basis for CFA’s $25 broadband overcharge? Kline doesn’t tell us in his press release. But a cursory look at the CFA study, which runs over 200 pages to create the veneer of serious scholarship, shows that the number derives from an arbitrary reduction in broadband prices needed to drop the EBITDA margin for Comcast’s broadband service to 40 percent, the same margin earned by Comcast’s cable video service. That’s it. Even if this method made sense for Comcast (and it does not), why Comcast’s margin differential reflects the overcharge for every ISP in America is never explained.
The CFA study offers a second justification for the $25 overcharge, in case the first one didn’t grab you:
In other words, removing the excess [broadband revenue] would split the surplus between producers and consumers. To put this in perspective, as shown in Figure V-6, cash flow per subscriber has increased by over $50 per month since the early days of high-speed data offerings by cable operators. A reduction of $25 per month would have split the increasing surplus between producers and consumers. (CFA Study at 170) (emphasis added).
This is not how overcharges are estimated.
Typically, an economist will find some competitive benchmark price that is not contaminated by the conduct being challenged (in the case, the exercise of alleged monopoly power). Or regressions are used to isolate the impact of the conduct on prices. But in arriving at its “EBITDA-based estimate of overcharges,” the CFA study simply waives its hands, solves for the broadband price that would yield the competitive EBITDA margin (with Comcast’s cable video as the benchmark), and appeals to an invented equity of splitting the surplus.
The basis for the $10 overcharge on by evil wireless operators is just as arbitrary. Here the CFA study computes the differential in EBITA per subscriber per month between AT&T and T-Mobile. Just as cable video served as the “competitive benchmark” in the broadband overcharge model, T-Mobile plays the “competitive benchmark” for wireless telephony. But there is no reason why all wireless carriers should earn the lowest EBITDA in the industry. T-Mobile might be operating at razor thin margins just to steal share and to overcome AT&T’s and Verizon’s inherent network advantages. CFA would have the reader believe that 100 percent of the margin differential is attributable to ill-gotten market power.
This is not to say that the broadband industry is perfectly competitively supplied in every neighborhood in America. Cable operators have been shown to drop prices significantly in the presence of overbuilder competition or fiber-based Gigabit entry, implying a certain degree of pricing power in select areas. But CFA’s claim that every broadband operator in America overcharges to the tune of $25 per month is completely unsubstantiated. Ditto for the claim on wireless overcharges.
CFA can do better than this. And so can tech reporters.
There’s no better way to announce your arrival than the unfurling of the Model X’s Falcon Wings.
“You can take your hands off the wheel now.”
“Holy crap. Can I take a picture?”
“Not if you plan to share it.”
So my friends, I’m left to describe this occasion with mere words.
The Model X was hurling towards a sharp bend on the Dulles Toll Road at around 55 miles per hour (mph). I test drove the most powerful package, the P90D, which goes from 0 to 60 mph in 3.2 seconds with the “Ludicrous” option. As we neared the turn, I lifted my hands off the steering wheel, and I placed my life in the hands of Autopilot.
What happened next was nothing short of miraculous: The car hugged the dotted lane separator, curving safely away from the barrier. I felt like a kid taking his first ride on a roller coaster.
The next trick was the lane change.
“Hit the left turn signal.”
And like that, the Model X patiently waited for traffic to pass on my left, and then changed lanes. Mind blowing.
“Now come back into the right lane, and turn on the right turn signal.”
I obliged but nothing happened. Why didn’t we move onto the shoulder? Because the Autopilot can tell the difference between a dashed line and a solid line. It will never cross the latter. The car has eyes and it processes everything it sees, including speed-limit signs.
Autopilot is a $2,500 upgrade on the base model. It includes an autoparking feature that was equally impressive. The Model X spots a parking space, turns away from the space, and then backs in seamlessly. No mortal could execute this maneuver, at least not this quickly.
Worried about the Falcon doors opening up on the car next door? Yeah, Tesla thought of that too. If you park too close to a neighbor—indeed, so close that the front doors won’t open—the rear doors make a snap calculation: If rear exit is impossible, the doors won’t open; if rear exit is possible, then the Falcon doors will adjust their upward trajectory to avoid hitting the car next door. These guys have thought of everything.
The base unit of the Model X, called the 60D, starts at $74,000. It has the identical power as the next model up (the 75D), but travels 37 fewer miles on a full charge (200 versus 237 miles). The salesman confessed that the 200-mile estimate for the 60D is the distance under ideal circumstances, and that I should expect something closer to 167 miles per full charge under normal circumstances. To get the extra 37 miles of capacity for the 75D, Tesla wants an additional $9,000, which works out to $243 per extra mile of capacity.
But here’s the kicker: If you get the 60D, and then regret your decision, Tesla will upgrade the car for $500 plus the $9,000 upgrade package. And if you’re worried about resale values, Tesla will peel off the 60D decal and replace it with a 75D. (I’m inclined to get the 60D and incur the $500 penalty if I underestimated my demand for distance.) The reason why the upgrade is so cheap is that Tesla doesn’t have to change any hardware; instead, it simply unlocks the last ten percent of capacity on the same battery with a software change.
For my purposes, the chance of going over 167-200 miles in a day is remote. I’ll charge the battery nightly at my home. Unlike the old iPhone batteries, charging the Tesla battery daily does no harm, even when the battery is not fully discharged when the charging begins.
For longer trips, the Model X’s navigation system plans the route and displays the Tesla charging stations along the way. The national map in the store created the appearance that Tesla has blanketed the country. But when we plotted a trip from Tysons Corner to Baltimore, I was surprised to learn of only a single Tesla charging station (located in Bethesda) along the route. This is unacceptable, but the salesman assured me that another station was planned in Laurel. He also insisted that the Model X can charge at non-Tesla stations, so long as I buy a converter.
The default seat arrangement in the Model X is five seats, with three seats in the second row. Five of us (plus the salesman) tried to cram into a seven-seat configuration and it was a disaster. No one was willing to sit in the third row, and folks were even complaining about the second row. This is a design flaw: The center of the Model X is too small to cram a wall of three seats into the second row. The six-seat configuration is ideal, with two captain seats in the second row, opening up space in the second and third rows.
So how much will a decently decked out Model X 60D set you back? The version I designed has $12,000 of upgrades, including $1,500 for the pearl white multi-coat paint, $2,500 for the black nappa leather seating, $3,000 for the six-seat interior, $2,500 for Autopilot plus Autopark, and $2,500 for smart-air suspension (required for for the six-seat configuration). The smart-air suspension automatically lifts the car a few inches during heavy snowfall.
If a friend owns a Tesla, you can get a $1,000 rebate referral code. But that just defrays the $1,200 delivery charge. So you’re looking at about $86,000 all in.
Relative to other luxury SUVs, the Model X can be rationalized with the expected fuel savings (about $8.31 for every 100 miles driven at current gas prices), a $7,500 income tax credit, your time saved on refueling and oil changes, and your smaller carbon footprint. And when the Falcon wings rise in your driveway, I suspect a neighbor (or four) will rush to the Tesla dealership. And you can’t put a price on that!
Capital continues to flee digital infrastructure. The table below compares the first six months of 2016 with the same period in 2014–the last year in which ISPs were not subject to Title II regulation.
Aggregate capital expenditure (capex) declined by nearly $2.7 billion relative to the same period in 2014. While Title II can’t be blamed for all of the capex decline, it is reasonable to attribute some portion to the FCC’s draconian rules. After all, the rules (needlessly) bar ISPs from creating new revenue streams from content providers, and (needlessly) expose ISPs to price controls. Both measures truncate an ISP’s return on investment, which makes investment less attractive at the margin.
This is incredibly frustrating because net neutrality protections could have been achieved though an alternative source of legal authority (section 706). Mr. Wheeler took ISP investment for granted. Bad assumption.
The Federal Communications Commission’s Path to Populism: A Search for Relevancy in the Digital Age
July 11, 2016
Event programming will begin at approximately 12:45 pm.
Fresh off the heels of a scathing critique by Judge Williams of the economic analysis in its Open Internet Order, the FCC needs to avoid any appearance of further “economics-free” rulemaking. In a mind-numbingly complicated and extended proceeding to decide how to retain the agency’s foothold over business data services (BDS), the FCC released three economic statements from its staff this week that purportedly justify the draconian actions it wants to take in that space. The agency even went so far as to reveal that it subjected its core economic analysis to peer review. A boon for economics, right?
Not exactly. The FCC is merely paying lip service to infusing its decision-making with economic analysis. As revealed in the peer review, the flaws in the underlying economic work that undergirds the proposed regulation of BDS (previously called “special access” services) are potentially fatal, rendering the analysis useless as the basis for the agency’s proposed regulations.
At least one party directly affected by the agency’s move has already expressed outrage over the broken process by which the FCC subjected its economist’s work to peer review, including the FCC’s decision not to release the peer review responses until comments were due. Rather than piling onto the FCC’s process errors, I’ll focus on the merits of the FCC’s economic support.
By way of a quick history, the FCC retained an external economist, Dr. Mark Rysman, Professor of Economics at Boston University, to establish “direct evidence of market power” in the supply of BDS in its Further Notice of Proposed Rulemaking (FNPRM). A finding of market power is critical to the FCC’s endgame, as the agency cannot compel a BDS provider to share its facilities at cost-based rates with a rival, nor subject its retail rates to price-cap regulation, in the presence of competition.
But as George Ford so ably points out, Rysman’s paper, which appears as Appendix B to the FRNPM, “says nothing about market power.” That BDS markets served by a single BDS provider exhibit greater prices on average than markets served by two providers does not rule out the possibility that BDS providers earn zero profits in so-called monopoly markets—that is, they generate revenues just high enough to cover their fixed costs. Moreover, Rysman’s estimates of the difference in prices between monopoly and competitive markets are not headline-grabbing: Before considering any critiques of his methodology, competition allegedly brings a modest price reduction of roughly ten and three percent for DS3 lines and DS1 lines, respectively.
The FCC sought peer review of the Rysman study by Andrew Sweeting, Associate Professor of Economics at University of Maryland, and Tommaso Valletti, Professor of Economics at Imperial College London. Interestingly, neither reviewer seems to be persuaded by Rysman’s empirical analysis.
Both reviewers criticized Rysman for relying on a cross-sectional database of buildings in 2013, rather than on panel data, which would contain prices for each building over a span of years. Sweeting notes that an “obvious concern” of Rysman’s limited dataset is the possible presence of “unobservable differences across customers that are correlated with both prices and competition by using census tract or county fixed effects,” thereby potentially contributing to biased estimates of Rysman’s competition variable. Rysman’s estimated price effects from competition could be upwardly biased, for example, if as Sweeting notes, cable providers “might be particularly good at picking off customers who want fancier services from the ILEC, so that in locations with [cable] competition, ILECs are left serving customers who are purchasing relatively cheap products. . . .” Sweeting concludes that “most economists would regard within-customer-over-time changes in prices, that could have been identified and estimated with panel data, as more compelling.”
In addition to the problems relating to the cross-sectional nature of his data, Sweeting identifies deficiencies in Rysman’s econometric model. For example, Sweeting notes that many of the contracts in Rysman’s dataset “are likely to have been negotiated some time prior to 2013, when local competition may have been different.” Sweeting notes that there has been significant growth in fiber-based cable networks since 2013, which suggests that extant relationships between prices and competition might no longer hold in 2016. He faults Rysman for failing to control for contract terms, such as duration, additional equipment, or contract date, all of which could explain pricing variation for the same products. Sweeting is also critical of Rysman’s assumption that a lack of a competitor’s presence in the building indicates limited competition, even when a competitor exists in the same census block.
Valletti also addresses the limitations of Rysman’s cross-sectional dataset: “[T]he question remains whether it is still possible that unobserved factors that can affect prices (particular demand and supply characteristics) differ within the census tract, and could drive the entry of CPs.” If so, and if Rysman failed to control for the “endogeneity of competitive entry,” then his estimates for the competition variables could be biased.
But perhaps Valletti’s most devastating critique occurs on page five, when he ponders whether prices per building (Rysman’s dependent variable) are even capable of responding to competitive entry given that DS1 and DS3 prices are based on tariffs:
If a service is tariffed, which I understand is true, for instance, of DS1 and DS3 services, then that service must be generally available to all at the same price. I also understand that the carriers can and do under the tariff differentiate services based on geographic locations, and that under the tariff prices can also vary, for instance, with volume, term commitment, and quality of SLAs. But if one could control for all these factors, the prices should not change with the number of competitors, as the same conditions must be offered to everyone. So my main point here is to understand why – having controlled for all the “right” factors – competition should have a role for tariffed services. Else the interpretation of the regression results could be substantially different: if, say, regulated prices could not react at all to the number of competitors, then the present statistical findings are simply pointing to the spurious correlation that competitors seem to enter in particular buildings where particular contractual elements (not observed by the econometrician) are present.
Put differently, Rysman’s basic premise that DS1 and DS3 prices should be correlated with competitive circumstances makes no sense so long as prices are tariffed and available to all comers at the same terms.
In seeming disregard to these significant criticisms, the FCC presses forward with its radical proposal, which would subject both telcos (incumbents) and facilities-based entrants (cable companies) to price controls. None of the economic statements released by the staff this week credibly addresses the critical errors reviewed here. Peer review is great in theory, but if doesn’t cause the Commission to alter its approach, then what good is it?
The Federal Trade Commission (FTC) can’t bring a suit against every merger that threatens competition. Given its scarce resources–it has limited staff and budgets for external experts–the agency has to pick its battles carefully. Among the set of the mergers that would raise prices, how should it choose?
There are some obvious filters. For example, the FTC should not pursue cases in dying industries: As dinosaurs are toppled by new technologies, the opportunity to preserve consumer welfare is limited. Nor should the FTC pursue cases where the victims are large, sophisticated buyers (think Fortune 100 firms).
Consider a subtle filter: The FTC should not pursue cases where the industry is characterized by large negative externalities. Before turning to paper-based office suppliers, implicated by the Staples-Office Depot merger, let’s focus on toxic chemicals to drive the point home.
Two tanneries are situated along a river, spewing pollution into the water like nobody’s business in proportion to the output of the firms. The tanneries are tired of competing, and decide to merge. The FTC hires a world-class economist from Berkeley, who estimates that the merger will lead to monopoly prices. Should the FTC block it?
There’s no doubt the merger is anticompetitive. But as any good economist knows, relative to the competitive equilibrium, a monopolist produces at levels closer to the socially optimal output in the presence of negative externalities. We want pollution to come down. An efficient way to achieve this objective is to permit the two firms to merge and raise prices, thereby discouraging consumption of leather-based products. (We should also enforce the relevant environmental laws, but that is outside the scope of the antitrust agencies.)
Now back to paper-based office products. Due to environmental concerns, and due to a need to quickly incorporate materials into work product, I prefer to read documents on my screen. But my (generally older) partners can’t print fast enough. Some appear to print every email. Any why not? The cost to the partner of printing an email is zero. And the cost to the firm is near zero.
If that doesn’t annoy you, consider these stats from the Paperless Project:
- Americans consume more paper per capita – upwards of 500 lbs. annually – than anyone else on earth. On average, a person in the United States uses more than 700 pounds of paper every year.
- The United States uses approximately 68 million trees each year to produce paper and paper products.
- The average office worker continues to use a staggering 10,000 sheets of copy paper every year.
- Discarded paper is a major component of many landfill sites, accounting for about 35% by weight of municipal solid waste.
- Pulp and paper is the third largest industrial polluter to air, water and land in both Canada and the United States, and releases well over 100 million kg of toxic pollution each year.
- 40% of the world’s industrial logging goes into making paper, and this is expected to reach 50% in the near future
Which brings me to the FTC’s dogged pursuit of Staples-Office Depot. I am skeptical of the merits of the FTC’s claims regarding anticompetitive effects. Although the price discipline of Amazon is hard to detect looking backwards, on a going-forward basis, Amazon Business, which hit $1 billion in sales in its first year and is growing at 20 percent per month, will likely displace these two dinosaurs in a matter of years.
But let’s grant the FTC its (unpersuasive) theory of harm. Why in the world should we encourage competitive pricing of paper-based office products, given the horrific things paper does to the environment? A monopoly provider of paper-based office products will raise the price of printing emails, which is a good thing in my book.
For now, environmentalists can only pray that Amazon chases the dinosaurs out of business, and then raises prices to monopoly levels. Either that or pray that Millennials have different habits when it comes to printing emails in the office.
For industries characterized by negative externalities, the FTC should think twice before stopping mergers to monopoly. Did I mention that alcohol consumption is associated with violence? Of course, no merging beer producer would make this argument before the antitrust agencies. But a snarky economist with no skin in the game just might.
Some economic thoughts are just too complicated for Twitter. I need about 560 words to vent properly on the merger conditions that the FCC imposed on Charter-TWC. So sit back.
The first condition that deserves economic scrutiny is the FCC’s seven-year zero-price regulation on interconnection. For seven years, Netflix will make zero contribution to Charter’s recovery of broadband infrastructure costs. Although economists object on efficiency grounds to placing 100 percent of cost recovery on the more price-sensitive side (broadband users) of this two-sided market, at least this condition bears some resemblance to the FCC’s theory of harm–namely, that the combination of Charter and TWC would increase the merged firm’s pricing power vis-a-vis national content providers such as Netflix. But if you really believe this, as my liberal friends do, then why not just block the merger? Why approve a merger that facilitates the exercise of market power only to regulate that exercise out of existence?
That Charter offered up this condition to see the deal though, or that Charter had not previously charged for interconnection fees does not make this condition anything less than blatant price controls. Apparently, price controls are in vogue these days, with the FCC regulating the price of business broadband service at both the wholesale and retail level. But to an economist, this stuff is hard to swallow. Consumers hate higher prices, but high prices are also a signal to broadband rivals to deploy duplicative networks. If you think there aren’t tradeoffs, check out what’s going on across the pond, where price regulation (in the form of mandatory unbundling) and the absence of facilities-based competition go hand in hand.
The second condition that riled me up was the obligation for Charter to invade a rival cable operator’s territory. This requirement is misguided for at least two reasons. First, the merger of Charter and TWC does not reduce competition in the provision of broadband service to any household in America, as the two cable operators serve non-overlapping territories. This condition, like so many other FCC merger-related requirements in the past, is completely divorced from the competitive harms raised by the merger. It is yet another example of the FCC using the merger review process as a means to regulate the industry.
Second, if Charter satisfies this overbuild requirement in the territory of a cable operator with fewer broadband subscribers, then the requirement perversely increases concentration at the national level.
To see how, consider the following illustrative example: Charter invades the territories of Cox, Cablevision, and other smaller cable operators, shifting three percentage points of market share (about 1.3 million subscribers with speeds of 10 Mbps down or faster) towards Charter. (The market share figures are from a January 26, 2016 story in Ars Technica.)
|Provider||Subs Post Merger||Share Squared||Subs Post Invasion||Share Squared|
|Wide Open West||1%||1||1%||1|
As the table shows, if Charters takes subscribers from smaller cable operators, then nationwide concentration (or HHI) increases from 1,732 to 1,860. This is just one illustration, and the result would hold so long as Charter invades any cable operator’s territory save Comcast.
Recall the FCC’s theory of harm is that the merger would permit Charter to exercise market power vis-a-vis Netflix. By perversely increasing nationwide concentration, the FCC’s overbuild requirement will have strengthened Charter’s pricing power!
There. I’m done.