Methodenstreit: A Reply to a Question from Greg O’Brien @ Cartt.ca about Media Concentration Research Methods

I have changed this post since putting it up last Wednesday (May 23, 2013). I have not done so substantively. Indeed, I have left all the data and main claims as they were.

What I have done, however, is remove some of the snark at the top and the bottom that I directed at Greg O’Brien at the outset. It’s unnecessary, and as a few colleagues, friends and others with my interests close to heart have kindly suggested, we need more civility in the internets, not less (see Blayne Haggard’s thoughts here).

Btw, the picture in Blayne’s post of a guy pounding away at a keyboard struck a chord; Kristina, my wife, will nod disapprovingly for sure; and its effect would be even greater still if you put five more words at the end of the word bubble: “about telecom, media, internet concentration”. I’ll think about that.

The revised version follows. A link to the original is here.

Last Friday afternoon, just as I was settling in for the first long holiday weekend, Greg O’Brien, sent me an email asking about media concentration research methods. Greg is the founding publisher and lead writer over at Cartt.ca — an industry trade paper that serves the telecom, media and internet industries here in Canada. The question is an important one and so I began to sketch out a reply.

I was advised, however, that it would be best to wait. The final replies to the Bell Astral hearings had yet to be submitted and, thus, addressing questions of methodology directly bearing on the hearings in public was out of bounds until the proceedings closed. No need to tip your hat to others about what you’re thinking. It was another in a long string of moments when my ‘academic’ persona tugged hard to break free of the short leash imposed on experts appearing before the CRTC.

The advice I got was superb. My advisors were dead right and I was wrong. While my inclination is always to just reply immediately and as fulsomely as I can, that is not always the smartest thing to do. Ask any journalists who knows me, or anybody for that matter, and they will tell you that I always freely share my ideas and don’t play coy.[1]

The advice I received was right. Bell was poking around in the same spot that O’Brien was and raised the same question that O’Brien does about the HHI thresholds used by “consumer groups” (they don’t refer to me or the consumer groups by name, nor do they speak of public interest groups) (See Bell Final Reply, page 2).

The core of his original email is below. My reply follows.

Date:       Fri, 17 May 2013 12:11:23 -0400

From:    “Greg O’Brien” <greg.obrien@cartt.ca>   Block Address

To:          “Dwayne Winseck'” <dwayne_winseck@carleton.ca>

Subject:   Research question

Hi Dwayne,

I just wanted to point out an issue I came across about the research on media concentration that is part of PIAC’s presentation to the Commission on Bell/Astral and a big part of the CMCRP, too. I did a little digging into Herfindahl-Hirschman Index (HHI), to figure out what it was and came across some info below that it looks like, from the links, the HHI index itself was changed or updated back in 2010 by the Federal Trade Commission and US Department of Justice.

Your research paper says the HHI and the thresholds of media concentration fall into three levels:

HHI < 1000 = Un-concentrated

HHI > 1000 but < 1,800 = Moderately Concentrated

HHI > 1,800 = Highly Concentrated

However, these links here, here and here seem to show that back in 2010, those HHI thresholds were altered so that:

HHI < 1500 = Unconcentrated

HHI > 1500 but < 2,500 = Moderately Concentrated

HHI > 2,500 = Highly Concentrated

That puts the HHI scores for many of the media mentioned in your report in the moderate or low range, I think.

To be honest with you, this is a bit too deep in the regulatory research weeds for a story in Cartt.ca. But I was wondering if you could explain the difference to me? Am I missing something? If not, does the research need to be altered/updated? Please let me know if I am wrong, or if we use different numbers for Canada.

Thanks,

Greg

My Response

Hi Greg,

Thanks for your inquiry.

Before I begin, please let me ask you to address specific questions about methodology or data to me since it was me that was hired to prepare evidence and write a brief in support of the public interest and consumer advocacy groups’ intervention opposing Bell’s revised bid to take-over Astral. My response is done solely in my capacity as a scholar and director of the Canadian Media Concentration Research (CMCR) project.

I wanted to send you my response earlier but was advised that it best to wait until the Bell Astral proceedings closed. Turns out, Bell was poking around in the same spot you were (see Bell Final Reply, page 2).

Let me also say, though, too, I was a bit hesitant about replying to you on account of the fact that the only other time you’ve spoken about my data, method or research at all was when you tweeted one of Bell’s allegations about my CBC revenue data at the very end of the reply phase for Bell Astral 1.0. That you tweeted about it then without asking me first about my views, and that your question now falls again at the very end of the reply phase, feels funny to me and I don’t quite like it. 

However, let me put that aside and try to answer your question because it is a good question.

I am aware of the new U.S. Department of Justice and Federal Trade Commission’s guidelines. David Ellis, who you also know, sent them to me earlier in the year. Please ask him about that.

Guidelines do change from time to time. While the U.S. replaced the revised 1997 version guidelines in 2010, there are a couple of reasons why they have not seeped into the scholarly literature and my research methodology specifically.

For one, when guidelines change academics will always take time to decide if the changes adopted are suitable to the field we’ve been working in. There has been a long-standing argument amongst scholars that the DOJ’s existing guidelines were already inappropriate for communication and that a ‘weightier’ test was required because of the freight communications media carry with respect to free speech, the free press, privacy, democracy, their role as public spaces vital to citizenship, many non-market attributes and other such concerns. I share such concerns (also see Eli Noam and C. Edwin Baker on this point; or Compaine and Goldstein for opposing points of view).

Second, the International Media Concentration Research (IMCR), of which I am a part, and which is, as you know, led by one of the world’s foremost experts in this area, Professor of Finance and Economics, Eli Noam at Columbia University (New York), set sail in 2008. Changing course midstream and with the larger debates just referred to still hanging in the air would have been unwise. The fact that the project has forty or so scholars studying long-term media concentration trends in as many countries around the world also suggests that you don’t change things just because things in the U.S. change. 

Of course, we must take heed of what the U.S. does, but it does not determine things everywhere else. Historical and international comparative references, amongst other things, are crucial too. You might also ask Professor Noam as well why the project stuck with the existing standards rather than change to the new ones midstream?

In short, one doesn’t jump from a set of standards over which there is already a lot of debate to looser ones without a great deal of thought. That said, one should not cling to outmoded ways of thinking either, and so I have been looking carefully at the new guidelines with an open mind.

Indeed, I brought the new DOJ/FTC guidelines with me to Montreal two weeks ago and was reading them in the run-up to and during Bell Astral 2.0.  As you will see on page 19, the guidelines not only set the thresholds at the higher levels you recite, but tell us what constitutes significant consolidation by pointing to the degree of change, i.e. transactions that move the dial 100 or more points in markets that are already modestly to highly concentrated.

Here’s what the new guidelines say with respect to transactions in:

Moderately Concentrated Markets: Mergers resulting in moderately concentrated markets that involve an increase in the HHI of more than 100 points potentially raise significant competitive concerns and often warrant scrutiny.

Highly Concentrated Markets: Mergers resulting in highly concentrated markets that involve an increase in the HHI of between 100 points and 200 points potentially raise significant competitive concerns and often warrant scrutiny. Mergers resulting in highly concentrated markets that involve an increase in the HHI of more than 200 points will be presumed to be likely to enhance market power (emphasis added, p. 19).

The chart below created on the basis of 2012 revenue data shows that, contrary to what you say in your email, none of the sectors implicated by the Bell Astral deal are at the low end of the new guidelines, except radio – as I never fail to mention.

More importantly, the Bell Astral transaction will move several sectors from moderately to highly concentrated status even by the looser standards of the new guidelines, i.e. an HHI score above 2,500, as the chart below illustrates. These sectors are:

• English-language Specialty and Pay TV (2525.2);
• French-language Specialty and Pay TV (4085.1);
• total Specialty and Pay TV (2512);
• the total French TV sector is already above 2,500 but would be pushed further to 2801.7.

It is also important to point out that the Competition Bureau in Canada does not use the HHI to set fixed benchmarks but rather to help it “to observe the relative change in concentration before and after a merger” (emphasis added, p. 19, fn 31). The Bureau does, however, state that when the four-firm concentration ratio (CR4) passes 65% it may step in to examine whether a merger “would likely . . . enhance market power, and thereby . . . lessen competition substantially” (p. 19, fn 31). You can look at the data in the chart below and reach your own conclusions on this point.

In addition, in terms of relative change, as the DOJ guidelines quoted above state, a transaction that moves the dial in moderately or highly concentrated markets by more than 100 – 200 points will “potentially raise significant competitive concerns and . . . be presumed to be likely to enhance market power” (emphasis added, p. 19). Based on the 2012 data shown in the chart that follows immediately below, here is a list of sectors implicated by Bell’s proposed take-over of Astral that would move the dial between 200 and 1200 points (change in HHI noted in parentheses):

• English-language Specialty and Pay TV (+416 points);
• French-language Specialty and Pay TV (+1215.1 points);
• total Specialty and Pay TV (+608.5 points);
• English-language Total TV (+236 points);
• French-language TV (+207.5 points);
• Total TV (+298 points);
• French language vertical integration between BDUs and broadcasters (+361 points).

Changes in Concentration Levels: Before and After Bell Astral, 2012 Revenues 

2012 Revenues	Bell Mrkt Share Before	After	CR4 Before	CR4 After	HHI Before	HHI After	CR4 2008	HHI 2008
Conv TV
ENG	30.7	30.9	90.7	90.9	2337.2	2347.2	96.1	2724.9
FR	0	0	95.1	95.1	4403.4	4403.4	94.5	4005.7
ENG + FR	22.6	22.8	82.9	83	2287.9	2293.5	86	2367.4
Spec & Pay TV
ENG	28	33.8	83.1	84.5	2109.2	2525.2	73.2	1543
FR	27.1	59.2	97.9	97.7	2870	4085.1	87	2755.1
ENG + FR	27.9	38	81.5	83.8	1925.7	2534.2	71.9	1451.7
Total TV
ENG	29.2	32.5	81.9	86.2	1891.2	2127.2	77	1762.2
FR	11.1	24.4	91.7	92.9	2594.2	2801.7	85.2	2389
ENG + FR	25.4	30.8	76.8	83.3	1691.5	1989.5	70.9	1486.7
Radio
ENG	9.8	21.9	51.6	59.6	822.6	1014.4	56.5	970.8
FR	0	27	84.1	84.1	2406.6	2406.6	90.1	2704.9
ENG + FR	7.9	23.2	53.4	62	825.3	1127.3	60	1047.2
VI & Network Media (2011)
ENG	31.3	31.8	83.2	84.2	1984.4	2014.9	N/A	N/A
FR	35.2	40.1	71.8	76.7	1872.1	2233.1	N/A	N/A

Also take note of the big changes not just by the standards of regulatory authorities but those of the recent historical past as well, i.e. since 2008, and notably for pay and specialty tv, total tv and radio.

As you can see, Greg, if this was purely an issue of methods and numbers, the CRTC should be very busy. And it is. This is why the Bell Astral 2.0 deal has received the critical attention it deserves, by the Commission and by people such as myself.

Finally, as I am sure you will have noted, I have updated and made the CMCR’s analysis of the 2012 data available on our website. I have the French- and English-language market 2012 data that corresponds to each of the sectors that we released the other day (radio, broadcast TV, specialty and pay TV, total TV), and for vertical integration between BDUs and broadcasting in both English- and French-language markets as well as for Canada

I really would be delighted to share all of our data sets with you under appropriate circumstances once the CRTC completes its deliberations on the current transaction. Doing this kind of research is not easy. There is much judgment involved and reams of data to be managed. I would like to trust that your question comes from a good place but I’m also acutely sensitive to the fact that there are many who toss barbs at researchers and, especially, critical ones, all the time. It really needs to stop, and if a full prof with tenure and a good salary can’t stand up to such attacks, who will? 

Ultimately, I always aim to improve my work and what I put out under the auspices of the CMCR. If you ever see anything in need of improvement, correction, qualification, etc., please let me know and I will, as is our standard practice, fix things while publicly acknowledging any errors we have made and your role in setting things aright.

Best wishes,

Dwayne

[1] To put a more scholarly spin on it, questions about research methods are difficult and often boring, but they can be really helpful when they clarify how we know what we know. They tend to be open ended (and wordy, too) which leads in many unforseen directions. German philosophers originally called such activities “methodenstreit”, or “methods dispute”, hence the title to this post. The notion of methods disputes is now common across philosophy and the social sciences and yes, that includes economics (see here and here).