addresses some of these limitations by allowing for endogenous entry, format differenti-
ation, and endogenous quality choices in certain formats, such as News/Talk, where they
allow stations to choose to be either high or low quality (to be high quality requires a
higher fixed cost). Using this more general framework, they estimate that there is still sub-
stantial excess entry and also excess investment in quality. The intuition for the latter is
quite similar to the logic for why there is excess entry. When total listening (or format
listening) is fairly inelastic, when a station invests in higher quality it is largely going to
be gaining listeners, and their associated advertising revenues, from other stations. This
type of business stealing, for quality improvement rather than entry, still implies that there
will be excessive investment from a social perspective. Overall they find that welfare
would be increased by reducing the number of both high- and low-quality stations by
around one-half in formats where there tends to be more than one station in a market.
One could imagine relaxing their remaining assumptions in several directions. For
example, a more realistic model of the advertising market might allow advertisers to ben-
efit when they can reach homogeneous audiences that can only be achieved when there
are many stations offering niche programming. Alternatively, one might model invest-
ment in programming quality in a richer way to capture the fact that there is a lot of per-
sistent heterogeneity in the audiences of stations in the same format. For example, in
markets with multiple news stations in Fall 2006, the leading station has, on average,
an audience that is two-and-a-half times as large as the second-ranked station, and in
82% of these markets the same station was the leader 10 years earlier (author’s calculation
using BIA data). Of course, a model of programming investments could also be used to
extend our understanding of how consolidation has affected listeners and advertisers.
72
Another interesting direction would be to analyze whether, even though there may
be excess entry into radio markets as a whole, certain formats are underserved. Under-
provision could result from listeners in these formats being less valued by advertisers; extra
fixed costs that make it too costly for any firms to serve them in small markets; or some
extra social value to the programming that stations and listeners may not internalize. As
mentioned in
Section 8.5.2, this question has arisen in relation to minorities, and I will
return to it in
Section 8.8 when considering the role of non-commercial broadcasters.
8.7. STRATEGIES FOR RETAINING LISTENERS
As mentioned many times already, the business model of commercial radio is to sell audi-
ences to advertisers. A challenge for this model is that listeners typically do not want to
listen to commercials, and so may seek to avoid them, potentially undermining the value
of a station’s advertising inventory (interested readers should read
Chapter 5 in this vol-
ume). The problem is even more severe for radio stations, especially music stations, than
72
For example, in the structural models of Jeziorski (2014b), Mooney (2010b), and Sweeting (2013), station
quality is treated as exogenous even if it can vary over time.
376
Handbook of Media Economics
for local television stations because radio listeners are usually less concerned about missing
the programming immediately following the break, so they may be more willing to
switch stations.
73
Indeed, when there are several stations playing similar music program-
ming, it is quite plausible that a listener who switches to avoid an ad will never return. If
multi-homing listeners are more valued than single-homing listeners, then station-
switching may be even more of a commercial problem. In this section, I will review evi-
dence on how widespread station-switching and commercial avoidance really are, before
considering strategies that stations may use to limit switching.
The traditional view of people in the industry is that avoidance of commercial breaks
is widespread. An Arbitron-sponsored survey by
Generali et al. (2011) found that 362
advertising agency executives reported that they believe that station audiences are, on
average, 32% lower during commercial breaks than in the minutes leading up to the
break, while station managers believe that they are 22% lower. Empirical evidence sup-
ported this view. For example,
Abernethy (1991) placed cassette recorders in the cars of a
sample of listeners and found that, on average, listeners switch stations 29 times per hour,
primarily in response to commercials.
Dick and McDowell (2003) estimate that in-car
listeners missed half of the commercials that they would hear if they did not switch sta-
tions. These statistics on in-car listeners are relevant because about 35% of listening takes
place in-car,
74
with an even higher proportion during the morning and afternoon dri-
vetime periods. In-car listeners have been estimated to be more than twice as likely to
switch during commercial breaks than listeners who are at home or at work.
75
On the other hand, Generali et al. (2011) claim that confidential PPM data shows that
there is much less ad avoidance.
76
They estimate that, on average, only 7% of the audi-
ence is lost during a commercial break, and only 4% during breaks that are 3 min or
shorter. However, the fall in audience is greater for listeners aged 18–34 (11%), which
is one of the demographics most valued by advertisers, and for music stations (12%,
vs. only 1% for “spoken word” stations).
77
Given the difference between these estimates
73
Television programmers try to exploit the fact that viewers will not want to miss the conclusion of a show
by scheduling more frequent breaks toward the end of a program (
Epstein, 1998).
74
This statistic is based on 2007 listening and data reported in Arbitron’s Persons Using Radio report
(
http://wargod.arbitron.com/scripts/ndb/ndbradio2.asp, accessed January 3, 2014). The number is sim-
ilar in earlier years.
75
MacFarland (1997, p. 89) reports that, based on a 1994 survey, 70% of in-car listeners switch at least once
during a commercial break compared with 41% and 29% of listeners who are at home or at work,
respectively.
76
As noted in Section 8.3 , PPMs measure any contact the wearer has with commercial radio (for example, at
the dentist’s office), rather than active listening. The earlier cited estimates are likely to be focused on
active listeners.
77
Unfortunately the study does not break out avoidance by location, but it does find that avoidance is par-
ticularly low during the morning drivetime period. This may be explained by many music stations car-
rying more talk programming during the morning drive than they do at other times of the day.
377
Radio
and both industry perceptions and earlier results, it seems clear that more analysis of how
much avoidance of commercials takes place, and who avoids them, would be valuable.
Based on the traditional perception,
Brydon (1994), an advertising consultant, argues
that “for advertisers, the key point is this: if, at the touch of a button, you can continue to
listen to that [music] for which you tuned in, why should you listen to something which
is imposing itself upon you, namely a commercial break?” He suggests that either stations
should play very short breaks which would not make switching worthwhile or stations
should “transmit breaks at universally agreed uniform times. Why tune to other stations if
it’s certain that they will be broadcasting commercials as well?” Unfortunately explicit
coordination between stations is both potentially a violation of the antitrust laws, and,
from a practical perspective, the fact that most stations in larger markets do not use
pre-recorded programming makes it difficult to coordinate by starting and ending com-
mercial breaks at precisely the same time. However, it is still plausible that stations might
try to align their commercial breaks as much as possible even if they cannot do so either
explicitly or perfectly.
In aggregate, stations do tend to play commercials at the same time. Based on
Sweeting (2009) , Figure 8.1 shows how many music stations played commercials each
minute between 5 and 6 p.m., using data from the first week of each month in 2001.
At least 15 times more stations play commercials at 5:23, 5:37, and 5:52 than play them
at 5:05. An obvious question is whether this pattern arises from a desire to coordinate on
Figure 8.1 Proportion of music stations playing commercials in a minute from 5 to 6 p.m. based on
data in December 2001. Source:
Sweeting (2009).
378
Handbook of Media Economics
playing commercials at the same time or some exogenous factor that makes playing com-
mercials at these times (and at similar times in other hours) especially attractive. Two such
factors can be identified from radio programming manuals and discussions with radio
programmers. First, listeners tend to switch on around the start of the hour, and they tend
to be particularly likely to switch stations if they hear a commercial as soon as they switch
on (
Keith, 1987). Therefore stations avoid playing commercials at the top of the hour.
Second, the way that Arbitron has traditionally measured station ratings means that a sta-
tion’s ratings may be increased if it keeps listeners over the quarter-hours (
Warren, 2001,
pp. 23–24).
78
Sweeting (2006, 2009) analyzes how far this pattern is driven by stations trying to
coordinate stations using his sample of airplay logs from contemporary music stations.
These logs identify not only the songs that are played, but also where commercial breaks
or promotions are placed between songs. One can then use information on when songs
started and the length of songs to estimate whether stations in a market were playing com-
mercials at the same time.
Sweeting (2006) presents a model where stations may either want to coordinate on
breaks, which follows from the logic above, or they may want to have commercials at
different times. As he shows, this could happen if stations, instead of trying to maximize
the audience of their commercials, try to maximize their average audience. It is not
unreasonable that this could be stations’ objective, as Arbitron reports measures of aver-
age audience size, not the audience of the commercials.
79
Average audience may increase
if stations play commercials at different times when coordination results in some listeners
switching off the radio. These models give different comparative statics for how the
degree of overlap should vary with observable market characteristics, such as the number
of stations, station ownership structure, and asymmetries in station listenership. In both
cases, the relationships should be stronger when listeners are more inclined to switch sta-
tions when they hear commercials, as they are during the afternoon drivetime, when
there are many in-car listeners and few of the music stations in Sweeting’s sample have
talk programming.
The empirical evidence lines up fairly consistently in favor of the version of the model
where stations prefer to play commercials at the same time even when exogenous factors
that make some times good for playing commercials in all markets are controlled for. This
suggests both that strategic factors contribute to the degree of overlap observed in the data
and, perhaps more interestingly from an economic perspective, that, despite the fact that
78
To be precise, a station is credited with a listener for a quarter-hour if that listener listens to the station for
at least 5 min during the quarter-hour. Therefore a listener who keeps listening between 5:25 and 5:35 can
count as much as one who listens from 5:15 to 5:45.
79
Arbitron continues to aggregate PPM data into average quarter-hour listening data for advertisers, even
though much finer data is collected.
379
Radio
advertisers cannot directly observe how many people listen to their commercials, stations
do appear to act to increase the audiences of the commercials, even if this might reduce
their average audience. One reason for this may be that local advertisers actually have a
good sense of how ads on different stations affect their sales,
80
and this effectiveness deter-
mines how willing they are to pay for future advertising time. If this is the case, stations
will want to maximize how many people actually hear their commercials.
Sweeting (2009) takes the analysis further by building a semi-structural model of the
commercial timing decisions, in the form of a coordination game, that allows for the per-
formance of counterfactuals.
81
In particular, he considers what would happen if each sta-
tion internalized the externality that its timing decisions imposes on other stations. For
example, if station A does not play its commercials at the same time as station B, then as
well as reducing the audience for its own commercials it will also reduce the audience of
B’s commercials. The estimates suggest that while the preference to coordinate commer-
cials has quite limited effects on the timing of commercials in equilibrium, in the sense
that non-strategic factors lead to the basic pattern shown in
Figure 8.1, commercials
would overlap almost perfectly if these externalities were internalized, at least during dri-
vetime hours. This also suggests another route through which ownership
consolidation—which should incentivize and facilitate more coordination—should be
profitable.
Of course, there are other strategies that stations may try to use to increase the effec-
tiveness of ads, although these have not received attention in the economics literature. An
interesting issue here is that ads that are effective, in the sense that listeners can recall the
product being advertised, may not be ads that listeners particularly like, creating a balan-
cing act for stations who want to both carry effective commercials and also encourage
listener loyalty. A set of studies have looked at different aspects of Clear Channel’s
“Less is More” strategy, which was introduced in 2004 in order to try to reduce audience
perceptions of advertising clutter.
82
The strategy had three components: (i) reducing the
total number of minutes of advertising; (ii) reducing the number of commercial breaks (or
“pods”) so that there were fewer interruptions to programming; and (iii) increasing the
number of shorter commercials (e.g., 30 s) at the expense of traditional 60-s commercials.
80
For example, if listeners have to make a telephone call to make a purchase, then it is quite common for an
advertiser to list different numbers on different stations so that they can monitor where their adverts are
most effective.
81
A motivation of doing so is that, in common with many discrete-choice games, the game has multiple
equilibria when strategic incentives are strong enough. The paper shows how the existence of multiple
equilibria, here in the form of stations coordinating on playing commercials at slightly different times in
different markets, helps to identify the strategic parameters.
82
“A Radio Giant Moves to Limit Commercials”, New York Times, July 19, 2004, http://www.nytimes.
com/2004/07/19/business/media/19adcol.html (accessed January 3, 2014).
380 Handbook of Media Economics
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.