Department of Social Sciences
Economics Working Papers
An Economics Perspective Ten Years After the
NAB Case
Brooks B. Hull
Economics Working Paper # 58
January 1989
-
,
ltm
The University of Michigan Dearborn
Dearborn, Michigan 48128-1491
(313) 593-5096
AN ECDNCMICS PERSPECrIVE TEN YEARS AFTER '!HE NAB C,AgE
Brooks B. Hull
Deparbnent of SOCial sciences
UM-Dea:dx>rn
Dea:dx>rn, MI
48128-1491
(313) 593-5305
Brooks B. Hull (Rl.D., University of Washi.n3ton, 1982) is A$istant
Professor of Econani.cs at the University of Michigan-Dea:dx>rn.
His
research interests include broadcast~ econanics, irrlustrial
organization, am envirornnental econanics.
Helpful ccmnents by Eugene
Silberberg, Keith Leffler, severin Borenstein,
are gratefully acknowledgEd.
am anonynous referees
'!hanks also to carroll Foster for
providi.rg codEd data am to Pat smith for prograIl'llli.rg assistance.
AN EXDNCMICS PERSPECl'IVE TEN YEARS AFTER '!HE NAB CASE
Abstract
'!he
u. s.
Justice Department brought suit against the National
Association of Broadcasters in 1979, charging that the NAB Television COde restricted the SUWly of advertising.
'!his paper examines
inplications of a collusive code, concluding that the code did not successfully serve this p.u:pose.
Television station sale prices were
no higher in markets with a high proportion of code subscriber stations.
stations in single station markets were no less likely to
subscribe to the code.
Finally, rates of retunl on broadcast finn ani
ne'blork stocks did not c:llan1e when the antitrust case was settled.
NAB
case
AN Ertion of stations 11I.1St subscribe.
'!he higher
NAB
case
the proportion of code subscribers, the higher will be profit for all stations in the market. Table 1 sunmarizes OIS regression results, the deperxient variable beirg the prices of ninety-six U. S. cxmnercial television stations sold between January 1976 am the code's suspension in March 1982.
Prices are adjusted for different nonths of sale by the
gecmetric mean annual f.bxiy Aaa corporate bcn:l yield. r:Ihe variable AUDIENCE is the rnnnber of households in the A. C. Nielsen designated market area viewirg the station 9:00 am to midnight. CPI is the Bureau of labor statistics Consumer Price Irrlex.
CABlE is
the percent of households subscribirg to cable in each designated market area.
SINGlE is a dummy variable set
sirgle station markets.
to one for stations in
CDDE is the proportion of television
households viewirg code subscriber station programs. 3
Ccx:le membership
is recorded six nonths after the sale to acc::xJUIlt for buyer expectations.
Table 1 about here
r:Ihe significant
am
positive coefficient on the AUDIENCE
variable shows the :inp:>rtance of viewer households to station's profitability.
Likewise, the CPI variable shows that station price is
sensitive to inflation.
'!he CABlE variable is not significant.
r:Ihis
5
6 NAB
case
result sinply confinns that the audience size variable is accurately inco:rporating the effect of cable on a station I s audience. a lOOllOpOly television station gives no special advantage.
am
SINGIE
effect.
Possessing '!he CABlE
variables are anitted fran the seccni equation without
IIrportantly, the extent of television code subscription has no
significant inpact on station profit.
A smaller proportion of stations in single station
Inplication 2:
markets should be code subscribers than in markets with two or IOO:re
stations.
If it enforces collusive behavior, the code is unnec::esscrry in single stations markets. already lOOllOpOlies
am
IOOnqx:>ly behavior.
stations in single station markets are
do not need the code to enforce appropriate
In November 1980,
seventy-two percent of stations
in nulti-station markets were code subscribers.
By contrast, only
thirty-nine percent of stations in single station markets subscribed (st:.arrlard Rate
am
Data service, 1980).
Although. it
~
consistent with a collusive code, the
difference in code subscription between single
am
markets is caused by differences in audience size. be a dunmy deperrlent variable.
am
SINGIE (both
defined above).
nultiple station let code membership
'!he inieperrlent variables are AUDIENCE
Prabit analysis yields a coefficient
on AUDIENCE asynptotically significant at one percent but an insignificant coefficient on SINGIE.
'!he nodel predicts fifty-eight
NAB
peramt of the (DOE outcanes.
case
stations in sirgle station markets have
fewer viewers than stations in multi-station markets ani stations with fewer viewers are less likely to subscribe to the code. 4 '!he positive relationship between code subscription ani audience size probably shows how smaller stations have less reaSon to ~rt any pmlic service or
latbyin:J characteristics of the code.
Inplication 3:
'!he rates of return on television broadcast irrlustry
stocks should fall when code enforcement
ems.
Retmns on cc:moon stock in part reflect expectations about finn perfonnance.
If the television code significantly increased station
profits, i.np:>rtant events in the antitrust case should reduce broadcastin;J
CCITpaIly
stock retmns.
Inportantly, this measure should
detect successful code-enforced collusion atoorg the networks ani in regional advertisirg markets in addition to the local. markets analyzed in the previous inplications.
Of
k.
Quarterly
Business, 21, 77-96. (various years).
Washin;)ton, D. C.:
Broadcastin:r
Broadcasting
Publications, Inc. Bram, S. J. & Warner, J. B. perfonoance. By the mnnbers.
(1980).
JOlln'lal of Financial Ecxmani.cs, ,§, 205-258.
(1988, 25 April) •
Children's television.
Measurirg security price
Broadcasti.m, p. 14.
(1988, 2 May).
Broadcastim, p. 14.
14
NAB
DeGroot, M. H.
Probability am statistics.
(1975) •
Rea~,
case
Mass:
Aaiison-wesley• Fama, E. F.
(1976).
Fam:latioos of fi.nar¥::e.
Foster, C. B. & Hull, B. H.
York:
Basic Books.
An O~ in fantasylam:
(1986).
television cx:de as cartel.
New
UM-Deal:bo:rn
'!he NAB
Econanics WOrkirp
no. 41.
~,
Foomier, G. M.
(1985).
Nalprice CCiipetitial am the dissipation of
rents fran television regulation.
scnthenl Econanic Joomal,
51, 754-765.
Foomier, G. M.
&
Martin, D.
(1983).
I))es
government-restricted entry
produce market power? New evidence fran the market for television advertisirg.
Bell Joomal of Econanics, (Sprirg),
pp. 44-56.
levin, H. J.
(1964).
Econanic effects of broadcast liamsirg.
Joomal of Political
levin, H. J.
(1975) •
Econcmv,
72, 151-162.
Franchise values, merit progranmi.n;;J ani policy
options in television broadcastirg. (Eds.) ,
Mass. : !£Ny, M. R.
In R. caves & M. Roberts
Regulatirg the product (pp. 221-247). BallanJer Publishi.rg
(1983).
cambridge,
co.
'!he time-shi.ftirg use of haDe video :recorders.
Joomal of Broadcastim ani Electronic Media, 27, 263-268.
Linn, s. C. &
~l, J. J.
(1983).
An enpirical investigation of
the impact of 'antitakeaver' cnnet'dnents on CCIlIOOn stock prices. Joomal of Financial Econanics, 11, 361-399.
15
NAB
National Association of Broadcasters. New York:
(varioos isS'teS).
case
COde NerNs.
National Association of Broadcasters.
National. Association of Broadcasters. (22m 00.).
New York:
(1981).
'!he television cxxie
National Association of Broadcasters.
National Association of Broadcasters, COde Authority.
(1980).
F\n'ctions am procedures of the cxxie offices (Men¥:>rarxlum, 29 May).
washin;Jtat, D. C.
Noll, R., Peck, M., & ltt::QJwan, J. television regulatim.
owen,
(1973).
washirgt:at:
Eoonanic aspects of Brooki..rgs Institute.
B. M., Beebe, J. H., am Mannirg, W. G. Jr. econanics.
Iexin;)too., Mass.:
(1974).
Television
D. C. Heath & Co.
Park, R. E., Jdmson, L., & Fishman, B.
(1976) .
Project.in:r the growth
of television b:roadcast.in:r; Inplications for spectrum use (Ram Report R-1841-FCC) • SdrNert, W. G.
(1981).
regulation.
steiner, G. A.
santa Monica, california:
Ram Corp.
Usirg financial data to measure effects of
Joornal of law am Eoonanics, 24, 121-158.
(1963) •
'!he people look at television.
New York:
Alfred A. Ily~
techniques in
DeGroot (1975).
Fama
(1976), Brown
tests the null
'!he statistic
hypothesis that the amulative average excess retums are significantly
different fran zero.
'!he statistic has an asynptotically nonnal
am
distribution with mean zero
st:amani deviation one.
'!he Z statistic is calculated as the followinJ:
z = CAR
CAR
=
/ S (AR) ,
N (1/N) ~ CARj ,
j=l
5tCAR:j) = { ~
S(AR)
=[
T
[1 +
l/T +
CRmt-Rm)2 /
~ CRm!t-Rm) 2
]
}~,
1
T-2 / N(T-4) ]'5,
where N
= l1\llli:)er
Q=
j
l1\llli:)er
=a
finn,
of finDs, in this case nine, of trading days in the test pericxl, in this case 100,
18
NAB
case
Rjt = d:Jserved daily :retum en finn j 's stock,
Rmt A
~
=
= dJserved
daily :retum en the market,
estimated parameter,
= variance of residuals
fran OIB :regxessien for the base pericx:l
for finn j,
Rut =
average daily :retum on the market duriDJ the base pericx:l, an:i
T = IlUIliJer of t:.rad.in;J days in the base pericxl, in this case 253.
19
NAB
case
Table 1 Regression Results with statim sale Price
00NSTANr
AUDIENCE
equation 1
-19.7
0.19**
equation 2
-18.4**
0.19**
CABIE
axE
SINGlE
10.1 -0.04
-0.03
-1.11
CPI
-1.27
8.89**
** i.mi.cates significant at one percent.
n
= 96.
R-BAR,2
0.72 0.72
20
case
NAB
Footnotes For a diSClSSiat of trade associatiat antitrost law see
1.
wilcox an:l She};ilerd (1975, p. 160) an:l Asch (1983, g>. 214-17). In studyirq the televisicn imustry, rsvin (1964, 1975)
2.
station sale price.
uses
Falrnier an:l Martin (1983), Boyer am wirth
(1981), an:l Park, Johnson, an:l Fishman (1976) use station aCXXAmtin; FCC Network Irquiry Special staff (1980) uses both acxx:mrt:in1
data.
data an:l station sale price.
use cxmnercial time prices.
Besen (1976) an:l wirth an:l Wollert (1984) Nale of tllese authors :in=l\Xle the NAB or
its code in their IOOdels.
sources for the data include Broadcastim-cablecastim
3.
yeartxx>k, various years; COde News, various issues; A. C. Nielsen, am Co., Market IBypart SUnlnaries, various issues; Rates
am
IBta, various i.sS1:aes.
recorded at time of sale.
am
Spot Televisiat
unless otherwise stated, data are
19>rcpriate variables are adjusted to
Nove.ni:er 1979. 4. larger
If code subscriptiat in:::reases profit J:ut statioos with
au:li~
code causes
are m:>:re likely to subscribe, a problem arises.
am is a result of
higher profit.
Foster
'Dle
am Hull (1986)
address this problem by enployin) a dlmmy erdogerD.JS variable IOOdel.
'!he IOOdel yields results consistent with those :reported here.
21
22 NAB
5.
'!he firms are the American Broadcastirg CQIpanies, CBS
Inc., capital Cities carm.micatioos, COX cemu.micatial, Gannett,
Liberty
case
corp., Metranedia, storer, am
Taft Broadcastirg.