NAB Turkey of A Report

As reported at Consumers Union Hearusnow.org blog The National Association of Broadcasters has done its best to show that owning broadcast stations loses money. Unsurprisingly, they recommend relaxing local ownership rules to allow owners to chase the happy, mythical synergy rainbow that has proven such a winner for Clear Channel, Tribune, Viacom and growing list of companies that absorbed profitable businesses and turned them into failing operations ladden with debt.

We shall leave aside the absurdity of the NAB’s arguments for the moment to get to something even sillier, the absurdity of the NAB’s math. Not since fictional Fundamentalists supposedly redefined Pi as 3 has ideology so distorted the basic precepts of mathematics. Worse, these are not accidents or “fudging.” I count no fewer than two major errors in methodology or presentation per page as well as many major methodological errors that impact the paper overall.

How bad is this paper? It is so bad that you would expect it to appear in the “April Fools” edition of Econometrica. It is so bad that I would expect its author, Theresa J Ottina, to be banned for life from meetings of the American Economic Association. It is so bad that every professor of economics and statistical analysis should download it and give it to their class as a final exam question to see if the class can spot all the errors as a kind of economics “Where’s Waldo” of mistakes, gaffes, and deceits. It is such a botched attempt at a lie by statistical analysis that I have half a mind to file a complaint with the FCC requesting they sanction NAB and Ms. Ottina for violating the FCC’s requirement that submissions reflect an honest effort to provide true information (a certification NAB made in its filing).

What makes it so bad? And why does the NAB submit such a piece of obvious crap?

See below . . . .

Yes, the people who brought you “Low Power FM will Blow Up Your Radio” and “Unlicensed In The Broadcast White Spaces: A Tawdry Tale of Spurious Emissions, Desensitized DTV Tuners, and Unrequited Reception” have submitted yet another whacky paper where the laws of reality take a holiday. As part of their 139 page comments in the FCC’s current inquiry on media ownership, the NAB submitted a study (Attachment J in this humungous PDF file) called “The Declining Financial Position of Television In Medium and Small Markets.” The study purports to show that station revenue keeps declining and therefore the FCC better relax local ownership rules quickly or free over the air television will die a miserable, impoverished death.

NAB, you understand, is pretty much a one note johnny when it comes to its arguments. Because Congress and the Supreme Court have recognized the value in maintaining a free over the air television network, the NAB always claims it needs regulatory goodies to avoid going bankrupt and killing free over the air TV. (Please note that Congress and the Supreme Court recognize the value of locally-owned free TV that provides genuine local news and information so that the citizenry can govern themselves, not just free over the air TV for its own sake. See Turner Broadcasting V. FCC and the sequel, Turner II –Justice Kennedy Strikes Back!).

We’ll pretend, for the sake of agrument, that allowing further consolidation would help stations become more profitable (although the recent experiences of Clear Channel and Tribune should raise some healthy skepticism in this regard). Instead, for those of us that like real math and love a good joke, let’s take a look at how the NAB reached its conclusions. After all, as Hearusnow observed, most broadcast television outlets operate at about a 40% or higher profit margin. But Hearusnow did not get nearly enough into the details.

Lets start with the methodology. First, the NAB selected three years to study. 1997, 2001, and 2003.

O.K., why these years? It’s not exactly a linear pattern. Heck, it skips 1999, as well as skipping every even year.

The NAB, in its comments, observes that it selected these years because: “None of these years involved a national election or the Summer Olympics to avoid the sometimes inconsistent impact of advertising associated with these events.”

That still doesn’t explain skipping 1999. Except, of course, for the fact that 1999 was an enormously profitable year for advertising. As some of you may recall, we were having an economic boom then, with lots of folks tryng to advertise, and businesses making money hand over fist.

So what the study does is select the three least profitable years it can find as a means of skewing the data. It then presents these three points as if they are a linear trend. This is how freakin’ high schoolers cheat on their friggin’ lab reports! If the graph needs to show a linear progression, take the three most favorable data points and draw a line between them.

To the extent the NAB is trying to manufacure a plausible excuse for this shameless effort to warp the data, it is hoping the reader once took statitstics and remembers it poorly. If you ever took a statistics course, you may recall that you generally expect to get a “bell curve” distribution. To increase relaibility, you occassionaly eliminate the highest and lowest values so that you have eliminated the “outliers” and get a more even distribution curve.

Here, however, the NAB has not eliminated the “outliers.” It has eliminated the most profitable years but kept the least profitable years. This skews the distribution curve artificially lower. Also, if elections and Olympics occur with regularity every two years, they are inherently part of the business cycle of broadcast stations. A cyclical spike in revenue is not an “inconsistency.” It is part of the equation. To eliminate predictable, cyclical revenue from calculation deliberately distorts the industry analysis.

But that’s not all. Here is a list of the other major errors in the study as a whole:

1) The study fails to identify costs associated with the DTV transition. These are one-time costs that do not impact the long-term profitability of a station. But because they occur only in the later years of the time studied by the NAB, they appear to depress profits over time.

2) The study excludes those markets in which the NAB did not have data on the highest rated station and the lowest rated station. But it presents the data as clusters of 25 markets as if it represented a general trend. (In other words, it pulls the same crap as it did on years of study. It takes non-representative data and tries to pass it off as a trend.)

3) Worse, according to figure 1, the number of stations reporting in a given cluster changes from sample year to sample year. The report provides no explanation for how this skews the data. We have no knowledge from the report how many markets reported for all three years or any explanation for how the study controlled for this shift.

4) Finally, the report excludes any market in which the highest rated station and the lowest rated station did not report data; the NAB does not explain how many other stations in the included market reported data. As a consequence, the study again skews the data toward the outliers and away from the reliability zone of the standard distribution curve.

5) The study examins four factors: “cash flow,” “pre-tax profit,” “network compensation,” and “cost of news coverage.” No explanation is given for these factors or how they relate to the question of overall profitability.

Bonus feature: There is a huge difference between “profitability” and whether the business is in fact profitable. A station could experience a drop in “profitability” while remaining an enormously profitable business if its annual rate of return dropped from, say 50% to 30%.

6) No distinction is made between network owned and operated (O&O) affiliates and independent affiliates. This is important because the study identifies the decline in station revenue as coming in large part from the decline in payments by networks to affiliates to carry programming. But if the network owns the O&O, this is a false loss, becuase it is the network no longer paying itself. So the rise in consolidation increases the decline in payments to affiliates in a market, because the total payment drops with each affiliate the network buys, and the NAB portrays this drop in payments in the market as a drop in profitability for the market as a whole.

7) The report uses self-reported data with no effort by the NAB to verify the accuracy of the reported data. Nor does it explain how it controls for the fact that overall responsiveness declines linearly for each year studied. (That is to say, fewer stations responded for each year in question.)

8) The report excludes the smallest 50 markets. Yet the theory of the report (and the way presented) is that it is the smallest markets that are suffering most. This might be forgiven if the other markets reported supported a correlation between market size and decline in profitability (as claimed by the executive summary and conclusion). But they don’t. Even based on the NAB’s skewed data, there is no discernible correlation between market size and profitability.

9) The report breaks out “highest rated stations” and “lowest rated stations” for some values, but explicitly averages highest rated stations and lowest rated stations for other values. No explanation is given for why this makes sense (probably because it doesn’t).

Even with all this fudging, the report fails to make the case it wants to make. In markets 51-75, the highest rated stations saw increases in profitability. In markets 76-100, both the highest rated stations and the lowest rated stations saw increases in profitability. In the smallest markets studied, markets 151-175, the largest stations saw a huge increase in profitability (155.5%). Only in two market segments, the 101-125 (which also experienced the highest differential in the number of stations reporting between years) and 126-150 was the report able to manufacture an apparent loss in profitability for all stations.

In other words, even the NAB report supports maintaining the “four station” rule, which prevents an entity from owning more than one of the top four stations in a market. According to the NAB, the top four stations not only remain profitable, they have, on average, increased profitability.

O.K., this is fun for me, but why would the NAB submit such obvious crap? If someone like me, who has only an undergraduate knowledge of statistical analysis can find more than seven methodological errors without breaking a sweat, how does the NAB hope to persuade an expert agency like the FCC?

As I observed recently, much industry “research” submitted in these filings is crap designed to play on the ignorance of the audience. Yes, the FCC has staff that are supposed to sort the crap from the non-crap. But the reality is that decisions are often driven politically the justified by pointing to evidence in the record that supports the political outcome. FCC Commissioners and members of Congress, generalists with a great many other areas of concern, can hardly be expected to go over this report with the same critical eye I just applied. And if you are a die-hard deregulator looking for some research to justify your decision, you don’t need to know the details. You just need to know that your buddies at the NAB have given you something that — on the surface at least — looks like a real report using actual math.

So the NAB just tries to make sure it has some crap in the record that looks plausible. It depends on the fact that no one in the public interest community or at the decisionmaking level is going to slog through a gajillion pages of supposed studies with supposed math — or would understand it if they did. So it just has to look like a serious study, thinks the NAB, but doesn’t need substance. You know, like in those war movies where the clever soldiers use cardboard tanks to deceive the enemy into thinking they are outnumbered and surrendering.

Happily for the public policy world, the day has long passed when the folks in the public interest community skipped the math parts and relied on unsupported appeals to basic principles. If you look at the studies submitted in the record to date, you will find that the public interest community has submitted a number of studies that actually conform to established principles for empirical research. Any serious study of the market leads to the same conclusion: the current media market has become over-concentrated. Not only would relaxation of existing rules be a serious mistake, but it would require significant divestitures to bring the media market to a state of health.

The NAB has opted to field an army of cardboard tanks to try to scare off the opposition. The public interest community has fielded an army of real tanks.

We will see who crushes whom.

Stay tuned . . . .