Misguided assumptions behind the CRTC’s broadband target (3)


How do you know?

This is an important question in many walks of life. In the natural sciences, it’s the most important question.

I’ve been doing some work with my son recently and had a chance to see that for myself. As a medical geneticist, Jordan devotes a lot of time to carrying out research and sharing his findings with other scientists in peer-reviewed journals. In his field, the standards of proof are extremely high, both because molecular medicine is so complex and because peoples’ lives are at stake.

I’d be exaggerating if I said lives were at stake in my classroom. Yet what I teach liberal arts students in a seminar about the Internet is based on a sense of respect for the same principle: you can’t write a research essay based on hearsay.

Students make bold assertions without giving a thought to why we should believe them. Of course, it’s a lot less work for the student writer to forget the “proof” and move on. But it’s well worth the effort on everyone’s part to encourage an understanding of when empirical evidence is important and when it isn’t.

Which brings us to the CRTC.

At the end of my previous post (part 2 of this mini-series), I was making invidious comparisons between the CRTC and its Washington counterpart, the FCC. I pointed to the great contrast between the FCC’s welcoming and open-handed approach to consumers, and the blank wall waiting for concerned citizens at the CRTC website. When you look at the FCC’s broadband page (here), you’re struck by two things. First, the FCC is falling all over itself to offer help to the visitor. Second, it’s doing so not by means of gimmicks or self-promotion, but by offering a wide range of real data and useful tools.

The CRTC’s framing of the Internet, on the other hand, is neither friendly nor informative. On its site that is. If, however, you look at what the Commission has said about the Internet in its published decisions, you discover a strange gap. Here e.g. is the opening paragraph of the public notice for the Son of UBB hearing, Review of billing practices for wholesale residential high-speed access services (Telecom Notice of Consultation CRTC 2011-77):

The Internet is a driver of innovation and the backbone of a modern economy. In recent years the way Canadians use the Internet has changed tremendously, in part because of the convergence of telecommunications and broadcasting services. It is vital that Canadians be able to access the Internet.

Who could disagree with this paean of praise for the online life?

But think for a minute about who the intended audience is for regulatory pronouncements, i.e. decisions, PNs, etc. – a very small community of licensees, lawyers, consultants, politicians and journalists. Why get on a soapbox to wax rhetorical for these guys when all they care about is the fine print and what it gives them?

On the other hand, what group of Canadians might need some reassuring that their being able to access the Internet is indeed “vital”? If any of our fellow citizens are worrying about access, it’s going to be those who currently don’t have access. Don’t get me wrong. Some of the regular participants in CRTC proceedings – PIAC comes to mind – are there precisely to voice their concerns about getting more Canadians online and making the Internet a better place for all.

But for your average schmo, what goes on at the CRTC remains a mystery, and the CRTC has done a great deal to ensure that “ordinary” Canadians remain mystified. Any Canadian who wants the CRTC’s help with any Internet-related matter is likely to find themselves at some stage on the FAQ page I highlighted in the previous post. The brushoff goes like this:

The CRTC does not regulate rates, quality of service issues or business practices of Internet service providers as they relate to retail customers. This is because there is enough competition in the market that retail customers can shop around for service packages.

The brushoff is one thing. What’s no less interesting is the Commission’s perfunctory rationale. We don’t need to help you by regulating your ISP’s behavior, because their behavior will be disciplined by competition in the market, which we know is adequate to ensure you’re treated fairly.

Here’s the problem. How the heck does the CRTC know there’s enough competition? I’ve long wondered whether the Commission actually performs any quantitative analysis that would give us some comfort about this assertion – especially since there’s strong prima facie evidence to the contrary (the OECD broadband data would be one kind of evidence). The CRTC’s 2010 Communications Monitoring Report (CMR) contains a reference that touched on the central issue: whether one or more firms in a market have enough market power to warrant being regulated. Here’s what we get (p.8 and footnote 9, the only reference in the 2010 CMR to market power):

In Telecom Decision 94-19, the Commission established a three-step process by which it could determine whether a telecommunications market is or is likely to become competitive for the purpose of considering forbearance applications.

The three steps consisted of (a) identifying the relevant market; (b) determining whether the applicant has market power with respect to the relevant market; and (c) determining whether, and to what extent, forbearance should be granted.

Two things to note. First, market power is not invoked here in connection with protecting consumers from abusive behavior on the part of licensed service providers. In other words, it’s not about whether the Commission should regulate a previously unregulated licensee or forborne service. It’s the opposite: its purpose is to help assess whether a regulated service should be forborne, on application by a licensee.

Where’s the homework?

Second, where’s the beef? Where’s the math? Wherever possible, regulation should be “transparent,” an unhappily modish concept which nevertheless points to a very important principle: show us the empirical work that led you to the conclusion retail broadband is competitive enough not to warrant being regulated on any basis (price, quality or firm behavior).

Economists (and others) have given us tools with which to assess such questions. One of the most widely applied is the Herfindahl-Hirschman Index (HHI), whose purpose in the hands of regulators is to assess market power and, if necessary, take measures to ensure firms are suitably disciplined. The HHI isn’t rocket science. You identify the market share of firms in competition and square those numbers. Regulators take these results and look for certain patterns such as whether one firm’s result is over a certain threshold or the HHI total has increased significantly over time. In the US, not only are the HHI, and similar indexes of concentration, used frequently by the FCC and DoJ. They’re also discussed as a matter of course in the media and telecom trades. The highly controversial AT&T takeover of T-Mobile has, not surprisingly, made references to the HHI popular reading.

HHI measures for US wireless industry, showing annual increases

associated with excessive market concentration (source: FCC, 2010)


In March, for example, widely followed GigaOm analyst Stacey Higginbotham wrote a pointed commentary on the AT&T deal (“AT&T, T-Mobile Merger: A Regulatory Quagmire?”) that included this analysis:

“Several sources in Washington D.C. were divided on the idea that the Department of Justice would approve AT&T buying T-Mobile. One source who has held multiple roles in telecommunications politics pointed out that this deal is a horizontal integration that can clearly show a decline in competitiveness, so it should be easy for the DoJ to stop, especially since the FCC last year clearly documented how concentrated wireless ownership was becoming, using an index the Department of Justice actually uses to determine the concentration of mobile wireless service providers: the Herfindahl-Hirschman Index.

“Under that index, the FCC found that the average HHI increased in 2008 relative to prior years. The weighted average of the HHIs was 2848 in 2008, an increase from 2674 in 2007. The weighted-average HHI has increased by nearly 700 since the FCC first calculated this metric in 2003. The industry has become more concentrated, leading to lowered capital expenditures and higher profits for the largest two players according to the report.”

The chart above is taken from a 2010 FCC study that fills 237 pages: Annual Report and Analysis of Competitive Market Conditions With Respect to Mobile Wireless, Including Commercial Mobile Services. The FCC is required to provide objective, quantifiable and publicly available measures of market power, rather than simply determining that a market is competitive by fiat. Moreover, it does so on an annual basis. Markets change, as the chart clearly illustrates.

Does the CRTC use the HHI? If not, how does it arrive at the conclusion that the broadband market is competitive enough? And does it apply such tools on a periodic basis to see if continuing market concentration warrants action this year that was not warranted last year? If there are calculations, why can’t we see them? Are they locked away in a filing cabinet marked Confidential to avoid causing irreparable harm to the incumbents? If the Commission’s decree – we know this market is competitive enough because folks can shop around for an ISP – were the thesis of an undergraduate essay, it would flunk.

Canada’s Internet problem is household income, not geography

The CRTC says Canada’s broadband problem is rural. Over and over again, we hear the mantra, which everyone in Ottawa has bought into, that the digitally dispossessed in this country are dispossessed by geography.

The CRTC also says that its policy challenges have nothing to do with digital literacy or affordability. That’s someone else’s problem, although the Chair has acknowledged there is a shortfall of 20-some percentage points between access and takeup in broadband. Were it not for too much geography, the Commission could rest on its laurels and just keep worrying about bandwidth hogs.

In the opening pages of its 2010 CMR, the Commission has the following comments about its data collection methods (pp. iii-iv, my emphasis):

The Commission collaborates with other government agencies and departments such as Statistics Canada and Industry Canada to minimize the reporting burden on the industry. The data collected for monitoring purposes is also used by Statistics Canada for its national system of accounts. Additional survey questions were added to meet Statistics Canada’s specific needs.

The Commission continues to work with Industry Canada to identify the availability of broadband Internet access service. The data, jointly collected, assists Industry Canada in the administration and monitoring of the $225 million broadband deployment initiative that was part of the federal government‟s economic incentive plan in 2009. Data collection forms are reviewed annually to ensure that only relevant data is collected.

It’s heartening to learn that Statistics Canada, one of the world’s most respected research agencies, has a role to play in Canada’s broadband policy, even if the price is that the industry is less forthcoming than it might otherwise be (reducing the “reporting burden” has a nice ring, like reducing air pollution or saving trees). Unfortunately, neither the CRTC nor Industry Canada seems remotely interested in what the Stats Can data are actually telling us.

On May 25, Stats Can released the latest figures from its ongoing Canadian Internet Use Survey (CIUS). While not comparable to previous surveys because of a change in methodology, they reflect the state of the residential Internet as of October and November 2010 (see The Daily). We learn 4 in 5 Canadian households (79%) are now online and most of those (about 96%) reported a high-speed connection. Thus, just over three-quarters of all Canadian households (76%) reported having home high-speed Internet access in 2010. Depending on whose figures you believe concerning how many Canadians have “access” to broadband, that would leave a shortfall in uptake of about 20 percentage points, as von Finckenstein has noted (the Stats Can data reported here are for households, not individuals).

On the other hand, those 20 points represent an average across different demographic groups (and other consumer segments), and policymakers have a responsibility to pull averages apart. Let’s see how Stats Can unpacks household income for Canadian onliners.

The chart above, concocted from the CIUS data, confirms in detail what we already knew about affordability in general. If your household income is $50,000 or above, i.e. in the top two quartiles, there’s at least a 90% chance you’re online at home (this covers both high-speed and dialup). If, on the other hand, you’re in the bottom quartile, earning $30,000 or less, there’s an almost 50% chance you’re not online at home, even on dialup. And that means the 25% of Canada’s households at the top of the economic food chain are almost twice as likely to have the Internet at home as the 25% at the bottom. (I note in passing that the survey instrument made available online leads into the income question as follows (p.6): “Now a question about your total household income. This information will be used to determine if Internet service is affordable to Canadians.”)

What are we to make of this distribution?

First, it shows the federal digital and broadband strategies in the poor light they deserve. The CRTC’s rejection of any responsibility for affordability runs directly counter to its grandstanding about the importance of access for all, unless “all” means everybody except the poorest members of our society. As for their counterparts at Industry Canada, former Minister Clement’s much-vaunted digital strategy was aimed at building the digital economy, meaning it was always aimed at business, never at consumers (the original strategy document never even mentions Internet access affordability). And the same minister and his boss produced a broadband program that was tightly framed in terms of the geography meme: the nation’s highest priority is overcoming geography, not social disadvantage.

(It’s probably too much to ask, but I wonder if any of the rural broadband projects that did get approved were vetted against community income levels.  Gentlemen farmers, Muskoka “cottagers” and other Canadians from the top income quartile don’t need any help getting online, but I’ve never heard any policy-related discussion about how such leakage could be prevented in the broadband program.)

So how does this compare with location of households? Stats Can has a very particular way of determining where you stand on the geographic food chain:

“A census metropolitan area (CMA) and a census agglomeration (CA) consist of one or more neighbouring municipalities situated around a major urban core. A CMA must have a total population of at least 100,000, of which 50,000 or more live in the urban core. A CA must have an urban core population of at least 10,000.”

Now look at this other chart (below), based on data from the same survey, this time showing the distribution of home access by location of households.


Until told otherwise, I’m going to assume that respondents located in neither a census metropolitan area (CMA) nor a census agglomeration (CA) can be considered as “rural” residents. In that case, the left-hand bar in the chart represents rural Canada, wherea the CA and CMA measures would together represent “urban” Canada (remember we’re dealing in households, not individuals). The chart shows that 71% of rural households are online, only 5 points away from those in a CA and 10 points away from those in a CMA.

These numbers don’t suggest to me there is a rural-urban gap that justifies focussing Canada’s entire broadband framework on geography, while we studiously ignore the social justice issues associated with low income. Even if Ottawa pays to get a connection into every hamlet in the land, we will never have universal Internet access – i.e. assurance that the vast majority of Canadians are actually online. It’s an embarrassment that our regulator, in keeping with years of smug back-patting, is still behaving like a PR agency and writing baloney like what follows (from the press release about the broadband target):

Despite Canada’s unique geography, 95% of households currently have access to Internet download speeds of at least 1.5 Mbps through telephone, cable or fixed-wireless networks. Over 80% of households already have access to download speeds of 5 Mbps or higher.

If you keep repeating this stuff often enough, you gradually lose your ability to see what’s nonsensical about it. If the Commission and officials at Industry can’t learn to talk straight about the painfully obvious, the only thing that’s going to change their thinking is progressive change in the international arena that eventually can’t be ignored – change that will come back to haunt the Conservatives where it really hurts, i.e. in our international competitiveness.

We’ve already got the data that show Canada is a broadband also-ran. And as of last Friday, we have the considerable moral authority of the Frank La Rue report for the United Nations: Report of the Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression. A typical headline from last Friday…

“United Nations Declares Internet Access a Basic Human Right”


What now, Ottawa?