FTC Staff Comment to the Hon. Lynn Adelman Concerning Wisconsin S.B. 47 to Repeal Auto Dealer Relevant Market Area Law

~1U1
a.st..t~~
~. IJ.IOia 60603
...,... Code 312353·1'1[,6
P'l.D£R.AL TRADt COMMISSION
Chiaro R.egoionaJ omce
.June 	19, 	1987
The 	Honorable 	Lynn Ade1mc.n
cnairman, 	~ud1ciary , 	Consumer
- 	, 	At fa i re 	Committee
WiQconein 	State 	Senate
State 	Capitol
P.O. 	Box 	7882
~dison, WI 	53707-7862
Dear 	Mr. 	Chairaans
The 	staf! 	of 	the 	Federal 	Trade 	Commission 	is 	pleased 	to 	have
the 	oPpoitunity 	to 	respond 	to 	your 	request 	for 	comments 	on 	Senate
Bill 	47. 	In 	essence, 	the 	bill 	would 	remove 	the 	portions 	of 	Wis.
Stat.: S 	218.01 	(1985-86) 	that 	provide 	existing 	retail 	automobile
dealers 	with 	adainistrative 	avenues 	for 	challenging 	the 	estab-
lishment 	or 	relocation 	of 	other 	dealerships 	within 	their 	relevant
market 	area I. 	These 	are 	hereinafter 	referred- to 	as 	the 	"RMA
laws·. 	As' we 	discuss 	below, 	we 	believe 	that 	the 	existing 	RMA
laws 	inhibit 	co.pet1tlon 	and, 	in 	80 	doing, 	may 	raise 	costs 	for.
consumers. 	We 	therefore 	support 	the 	passage 	of 	SB 	47. 	By
removing 	restrictione 	on 	competition, 	S5 	47 	may 	result 	in 	lower
automobile 	prices 	for 	consumers. 	We 	are 	also 	taking 	this
opportunity 	to 	comment 	on 	two 	other 	provisions 	of 	5 	218.01 	th~t
we 	believe 	also may 	raise 	consumer 	coste.
The 	Federal Trade 	Commission 	is 	charged 	by 	the 	United 	States
Congress 	with 	maintaining 2 com p etition 	and 	protecting 	consumer~
from 	restraints of 	trade. 	In 	accordance 	with 	this 	role, 	the
Commission 	and 	its 	staff 	submit 	co~ente upon 	request 	to 	federal,
state, 	and 	local 	governmental 	bodies 	to 	help 	assess 	the
competitive 	and 	consumer 	welfare 	implications 	of 	pending 	policy
issues.
1
2
These 	comments 	represent 	the 	views 	of 	the 	Cbicago 	Regional
Office 	and 	the 	Bureaus 	of 	Competition, 	Consumer 	PTotection,
and 	Economics 	of 	the 	Federal 	Trade 	Commission 	and 	do 	not
necessarily 	represent 	the 	views 	of 	the 	Federal 	Trade
Commission or 	any 	Incividual 	Commi8~ioner. 	The 	Fede~~l
~rade 	Commission, 	however, 	has 	voted 	to 	authorl%e 	the
sub~ia8ion of 	thesG 	comments.
See 	15 	U.S.C. 	S 	41 	et 	~eq.
--~.- -_., 	--- 	- 	--- 	---.-:, .,....-.......... -.. - 	.--. 	-- -

The 	Honorable 	Lynn 	Ade1~n
Poge 	2
Section 	5 	of 	the 	Federal 	Trade 	Commission 	Act 	prohibits
unfair 	methoda 	of 	competition, 	an~ 	unfair 	or 	deceptive 	aots 	or
practices. 	By 	enforcing 	this 	statutQ, 	the 	eo.aiaaion 	ataff 	has
gained 	substantial 	experience 	in 	analyzing 	tbe 	impact 	of 	various
~estralntB on 	competition 	and 	the 	costa 	and 	benefits 	to 	consumers
or 	such 	restraint.. 	In 	recent 	year., 	the 	Commission 	statt 	has
been 	involved 	1n 	a 	number 	ot 	issues 	specifically 	dealing 	with
retail 	automobile 	dealerships. 	In 	January 	1986, 	lor 	txamplQ, 	the
FTC's 	Bur~au of 	Economics 	published 	the 	results 	of 	a 	study 	of 	th)
REflect 	of 	State 	Entry 	R~gulation on 	R4tail 	Automobile 	~rkets.
COmmission 	staff 	has 	also 	gained 	experience 	relating 	to 	the
automobile 	industry 	through 	many 	investigations 	and 	litigated
easelS.
We 	believe 	that 	the 	current 	Wisconiin 	RMA 	la~ 	tends 	to
restrict 	competition 	unn~ces8arily. 	It 	require. 	motor 	vehicle
manufacturers 	an~ 	distributors 	who 	seek 	to 	establish 	or 	relocate
automobile 	dealerships 	to 	give 	written 	notice 	to 	e~isting dealers
Bellino 	vehicles 	of 	the 	same 	make 	within 	the 	relevant 	market
are a • 	~ (W is. 	S ta t • 	SS 	218. 01 ( 1) (r): 	( 2) (c) 2-. b ; 	3 (a ) 32; 	3 ( f) 	and
9 (a) (1985-86) ).' 	A 	-relevant 	market 	area - 	is 	de! ined 	as 	the 	area
within 	a 	lO-mile 	radius 	of 	the 	aite 	of 	an 	existing 	daalership 	or
the 	existing 	dealer's 	area 	of 	sales 	responsibility, 	whichever'1s
greater. 	ThQ 	dealers 	entitled 	to 	notification 	may, 	within 	thirty
days, 	file 	~ 	protest 	with 	the 	offll e 	of 	the 	transportation
commissioner 	alleging 	-good 	cause" 	for 	not 	permitting 	tha
proposed 	dealership. 	Upon 	rece~pt 	of 	a 	protest, 	the 	office 	of
the 	commissioner 	must 	schedule 	a 	hearing. 	If 	the 	matter 	is 	not
settled, 	a 	hearin9 	is 	held 	and 	a 	written 	opinion 	issued.
3 	We 	are 	enclosing 	a 	copy 	o~ 	this 	Staff 	Report 	(Attachment 	1)
for 	your 	info=~tion.
In 	oetermining 	if 	"good 	cause- 	t~iitl 	the 	state 	must
con~ider -the 	existing 	circuroEt~nces, including, 	b~t 	not
limited 	to:- 	(1) 	the 	£upply 	ana 	de~nd of 	vehicles 	of 	the
same 	mak_: 	(2) 	the 	n&ture 	of 	the 	exi~ting dealers' 	inVEst-
ment; 	(3) 	the 	effect 	on 	the 	retail 	motor 	vehicle 	~rket 	in
the 	area, 	(.() 	the 	public 	welfare; 	(5) 	the 	effect 	on
competition; 	(6) 	the 	qU&lity 	of 	54H\01ces 	being 	provi6e~ by
exi~ting de&lers; 	(7) 	the 	quantit.ies 	of 	vehicles 	en6 	~:-t6
pr~.ise6 to 	e~isting ~eale:s compared 	to 	those 	&ctu~lly
6elivereo1 	l.~d 	(8) 	t.he 	effect 	of 	a 	denial 	on 	~n.~ppl:ct.:-;t.
OHs. 	Stat. 	S 	218.01 (3) (!) (2)} • 	.

The 	Honorable 	Lynn 	Adelm.an
Page 	3
Senate 	Bill 	'7 	wilt 	B~nefit Consumers
SB 	(7 	would 	repeal 	all 	the 	portions 	of 	Wis. 	Stat. 	S 	218.01
dealing 	with 	the 	tights 	of 	exlat1ng 	dealers 	to 	notification,
prote8t 	and 	a 	hearing. 	We 	support 	this 	change. 	We 	believe 	that
sa 	47 	is 	likely 	to 	increase 	competition 	in 	retail 	automobile
markets 	in 	Wisconsin 	and 	thereby 	benefit 	consumere. 	Thi8
conclusion 	is 	based 	in 	large 	part 	on 	the 	FTC's 	Bureau 	of
Economics 	Staff 	ReportS 	which 	concluded 	that 	~ 	laws 	raieed
automobile 	price., 	on 	Bverage, 	by 	six 	percent.
Because 	RKA 	laws 	limit 	manufacturer.' 	freedom 	to 	ex~nd 	in
areas 	where 	automobiles 	are 	most 	in 	demand, 	the 	RMA 	laws 	tend 	to
increase 	the 	cost 	ot 	distributing 	automobiles 	and 	thus 	facilitate
escalation 	of 	automobile 	prices. 	Consumers 	may 	find 	fewer
opportunities 	to 	shop 	among 	dealerships 	for 	better 	prices 	and 	may
have 	tp 	travel 	farther 	and 	expend 	more 	time 	in 	the 	couree 	of
the! r. 	shopping.
5
6
Since 	the 	release 	of 	the 	Bureau 	of 	Economics 	(BE) 	Staff
Report, 	Wharton 	Econometric 	Forecasting 	Associates 	(WEfA),
at 	the 	r~uest of 	the 	National 	Auto 	Dealers 	~sociation, has
p~epare~ a 	report 	that 	reviews 	the 	Staff 	Report 	and
challenges 	its 	conclusion 	that 	prices 	of 	automobiles 	would
rise 	as 	a 	reEult 	of 	RMA 	laws 	in 	areas 	of 	growing 	popUlation.
The 	WEFA 	report 	also 	~iscus6es the 	theoretical 	arguments
that 	RMA 	laws 	are 	desirable. 	Attachment 	2 	is 	a 	Bureau 	of
Economics 	paper 	that 	responds 	to 	the 	*EFA 	report. 	WEFA's
speculation 	that 	the 	Staff 	Report 	has 	demon9trat~d only 	that
rap1~ly 9r~ing areaS 	have 	higher 	automobile 	prices 	rather
than 	that 	the 	price 	increases 	are 	a 	result 	of 	R!"..h 	laws 	is
shown 	to 	be 	unfoun6ed. 	The 	theoret~cal 	ArQUments 	that 	RKh
laws 	are 	desirable 	are 	also 	shown 	to 	be 	unfounded.
~~ese price-increases 	that 	result 	from 	RY~ lA~! 	~re 	most
likely 	to 	occur 	in 	counties 	or 	metropolitan 	areas 	exper-
iencing 	in=reases 	in 	popUlation 	(and 	the:~fore increases 	in
demand). 	The 	results 	of 	this 	FTC 	study 	~re cons~stent with
other 	studies 	showinS 	that 	increased 	prices 	result 	from 	R~~
laws. 	See 	tcKard, 	E.W., 	Jr., 	-The 	Eff~cts of 	State
Automo~ile Dealer 	Entry 	RegUlations 	on 	New 	cer 	Price&,·
Econo~ic Incuiry, 	Vol. 	XXIV, 	No. 	2 	(~pril 	1985), 	pp. 	223-~2,
and 	Smith, 	R.L. 	-Franchise 	Re;ulation: 	kn 	Economic 	~~alYEi5
of 	State 	~strictions on 	kutornobil~ Dist:ibution.- 	Jou:r~l
o! 	~w 	an: 	Econ~ics Vol. 	xxv 	(April 	1982), 	pp. 	125-57.

The 	Honorable 	Lynn 	Adelman
P~ge 	4
RMA 	l~ws 	may 	also 	provide 	op?Ortunities 	for 	existing 	dealers
to 	deter 	or 	~eley 	the 	entry 	of 	new 	competition 	by 	filing 	epuriou8
protests. 	The 	costs 	to 	existing 	dealers 	of 	such 	proteste 	are
probably 	small 	relative 	to 	the 	potential 	gains. 	A prote~t 	cen
Aelay 	the 	establishment 	of 	a 	new 	dealership 	by 	six 	months. 	Even
if 	only 	a 	small 	percentage 	of 	protests 	succeed, 	they 	may 	impose
eignificant 	costs 	on 	entrants 	and 	reduce 	the 	expected 	profits
from 	entry. 	In 	ad~ition, by 	interposing 	delay, 	an 	existing
dealer 	hae 	sn 	opportunity 	to 	negotiate 	some 	concession 	from 	the
manufacturer 	in 	settlement 	of 	the 	protest. 	The 	costs/of 	such
delays 	and 	concessions 	are 	ultimately 	PbBsed 	on 	to 	the
consumer.
We 	believe 	that 	exi~tin9 dealers 	are 	not 	likely 	to 	be
injured 	by 	repealing 	these 	RMA 	laws, 	except 	where 	the 	laws 	m~y
have 	insulated 	a 	dealer 	from 	competition 	and 	thereby 	permitted
the ~ealer to 	realize 	excessive 	profits. 	There 	is 	little 	basis
for 	present 	dealers' 	fears 	of 	being 	flooded 	with 	more 	dealers
th~n potential 	sales 	in 	a 	market 	can 	BU?port. 	It 	would 	not 	be 	in
the 	economic 	interest 	of 	either 	prospective-oealers 	or 	manu-
facturers 	to 	invest 	in 	such 	a 	market. 	Prospective 	dealers 	are
likely 	to 	give 	careful 	consideration 	to 	what 	the 	~rket will 	bear
in 	order 	to 	protect 	their 	investment 	in 	real 	estate, 	inventory
and 	personnel. 	~~nufecturers ~ill 	likewise 	avoid 	over-saturating
a 	market, 	because 	they 	have 	an 	interest 	in 	the 	financial 	well-
being 	of 	existing 	dealers. 	A manufacturer 	that 	treat! 	existing
dealers 	badly 	is 	likely 	to 	have 	difficulty 	finding 	new 	dealers
when 	the 	need 	~ri6es.
~~ 	laws 	do 	not 	appear 	to 	be 	necessary 	for 	the 	adequate
protection 	of 	dealer's 	rights. 	Their 	righte 	are 	protected 	by
exi~ting 	laws 	and 	the 	right 	to 	sue 	in 	the 	same 	fashion 	ae
franchisees 	in 	any 	other 	industry. 	Consequently, 	it 	is 	difficult
to 	see 	Why 	automobile 	dealer! 	require 	sp~cial tr6atroent 	under 	the
l~v.
Without 	the 	restraint 	that 	the 	RMA 	laws 	place 	on 	menu- •
facturers' 	ability 	to 	meet 	competition, 	manufacturers 	c~n 	be 	mc~e
responsive 	to 	consumer 	demands. 	It 	is 	also 	unli~ely thet
repealing 	RMA 	laws 	will 	reduce 	the 	quality 	of 	cealer 	service
because 	vigorous 	competition 	among 	manufacturers 	~ill 	ins~:e that
consu~ers receive 	the 	~ix of 	service 	quality, 	Quantity 	and 	price
they 	ce!ire. 	.

The 	Honorable 	Lynn 	Adelman
Pag_ 	5
Other 	Provisions 	Th~t May 	H~rm Con8umers
Two 	other 	provisions 	of 	S 	218.01 	~y 	also 	raise 	consumer
coste. 	One 	provision 	gives 	existing 	~ealers the 	right 	to 	60-cays
notice 	of 	a 	manufacturer's 	intent 	to 	terminate 	a 	franchise 	and
the 	right 	to 	an 	administrative 	hearing 	concerning 	the 	fairnes& 	of
the 	t~rruination. 	(wis. 	Stat. 	S 	218.01(2) (bd) 	an~ 	:218.01(3) (a)17
(1985-86». 	This 	notice 	and 	hearing 	provision, 	like 	the 	pro-
visions 	addressed 	by 	SB 	47, 	may 	~ecreaee 	manufacturers' 	ability
to 	adapt 	to 	changing 	conditions 	and 	thus 	hinder 	the 	establishment
and 	maintenance 	of 	efficient 	distribution 	systems. 	Kanufacturers
may 	also 	be 	th~arted from 	terminating 	inefficient 	dealers. 	The
result 	1s 	that 	consumers 	are 	denied 	the 	procoapetitive 	benefits
of 	a 	responsive 	distribution 	system and 	are 	forced 	to 	pay 	the
higher 	costs 	associated 	~ith 	an 	inefficient 	system. 	Without 	this
provision, 	dealers 	would 	remain 	protected 	from 	unfair 	termin-
ations 	by 	the 	terms 	of 	their 	contracts 	and 	their 	private 	right 	to
. sue, 	including 	the 	option 	to 	seek 	a 	preliminary 	injunction 	which
could 	maintain 	the 	status 	quo 	until 	a 	final 	judicial 	determin-
ation 	regarding 	the 	legality 	of 	the 	terreination.
The 	other 	provision 	requires 	dealers 	to 	clo~e on 	Sundays,
unless 	they 	close 	on 	Saturday! 	instead 	for 	religious 	reasons.
(Wis. 	Stat. 	S 	218.01(3) (a)21 	(1985-86». 	Studies 	have 	shown 	that
~ndato=y Sunday 	closing 	laws 	may 	harm 	consumers 	directly 	by
increasing 	the 	costs 	and 	inconve9ience 	of 	shopping, 	and
indirectly 	by 	increasing 	prices. 	For 	many 	people, 	the 	option 	to
shop 	on 	Sunday 	enhances 	the 	ease 	and 	convenience 	of 	b~ying a
car. 	Eliminating 	one 	of 	the 	weekend's 	two 	days 	from 	car 	shopping
forces 	potential 	buyers 	to 	make 	personal 	and 	economic 	.acri:ices
they 	would 	not 	othe:~ise have 	to 	make. 	If 	they 	cannot 	~hop on
Sunday, 	some 	pote~~ial 	buyers 	may 	find 	it 	neceesa:y 	to 	take 	t:rue
from 	work, 	give 	up 	some 	activities 	reserved 	for 	Sat~=day, or 	lose
leisure 	time 	on week~ights. 	Measuring 	the 	cost 	to 	consumers 	of
inconvenie~~e or 	los~ 	leisure 	is 	di~ficult, but 	the 	cost 	is
real. 	The 	costs 	of 	taking 	off 	work, 	or 	taking 	several 	wee~
nights 	to 	accomplish 	what 	can 	be 	done 	during 	one 	full 	day 	is 	more
easily 	measured 	in 	terms 	c~ 	lost 	wages, 	vacation 	tlme 	and
transport~tion and 	other 	costs.
i See, 	e.g., 	Carlson 	0 	Gieseke, 	~Price Search 	In 	a'P:06uct
Koartet,· 	Journal 	0: 	Consumer 	Resee.rch, 	March 	19S3, 	pp. 	357-
65, 	Morri~on « ~e.~n, 	in 	'Eou:e 	0: 	Operation 	P~strictionE
and 	Competition 	~~ng 	Ret~il Pirms,· 	Econ~ic :noui:v, 	Vol.
XXI, 	Jenuary 	19£3; 	Stigler, 	·7~~ 	EconorolC! 	ot 	:~:o=wetion,·
Journal 	of 	?o~1tica~ Econo~v, Jun. 	~961, 	pp. 	~13-2S.

t.
The 	Honorable 	Lynn 	Adelman
Page 	6
Increa8~d 	shopping 	costs 	may 	lead 	to 	increased 	automobile
prices. 	~s 	the 	co~t of 	shopping 	rise&, 	consuaers 	may 	~hop 	less
for 	automobiles. 	To 	the 	extent 	that 	consumers 	who 	shop 	less 	are
Jess 	informed 	about 	the 	market, 	they 	may 	be 	willing 	to 	pay 	higher
prices 	than 	vould 	be 	likely 	in 	a 	well-informed 	market.
Conclusion
In 	summary, 	we 	en~orse 	SB 	47, 	which 	would 	remove 	certain 	of
the 	restraints 	imposed 	by 	the 	present 	law 	on 	the 	retail 	auto-
mobile 	~rket. 	By 	removing 	these 	restraints, 	we 	believe
Wisconsin 	will 	be 	continuing 	its 	tradition 	of 	progre~aive
consumer 	legislation. 	We 	also 	encourage 	you 	to 	consider 	legis-
lation 	eliminating 	the 	other 	provisions 	discussed 	in 	this
letter. 	We 	appreciate 	this 	opportunity 	to 	provide 	our 	views 	on
5B 	47 	and 	related 	provisions.
Very 	truly 	your~,
},I~ 	r.~
William 	P. 	Golden
Actina 	Director
Chicago 	Regional 	Office
TTH: jd

6/03/87
RespODU to Wharton EcoDometric FOrtcastlD& Auoclatu' CommtDts
O. tbe Barca. or EcoDomlcs Study or Releuat Mark.~t Area Laws
Thu paper l responds to comments by VI'barton Econometric Forecasting
Associates C-WEFA-)2 concerning a 	1986 Bureau of 	Economics Staff Report
entitled 	lbe 	Effects 	of 	State 	Entry 	Regulation 	on 	Retail 	Automobile
Markets- (-Staff Report-). 	The Staff Report concluded 	on 	the 	basis of an
econometric 	study 	tbat 	state 	relevant 	market 	area 	laws 	rRMA 	laws-)
increase the retail prices of automobiles..' 	The WEFA comments suggest that
tbe 	Staff 	Report's 	conclusions 	are 	not 	valid 	because 	of 	errors 	in
specification 	of the 	price 	model. 	Below, we 	respond 	to Vt'EFA's comments
and, affirm our earlier conclusion 	that RMA laws increase retail automobile
prices in areas of positive population growt'h.
In addition to criticizing the Staff Report, WEFA argues that in theory
RMA laws may be desirable and that these laws will not be likely to impose
costs on society. 	Below, we respond to ViEFA's comments and conclude that
RMA laws are neither necessary nor desirable for the proper functioning of
the 	automobile 	dealer 	market. 	Instead, 	bocause 	of 	the 	price 	effects
discussed above, we conclude that these laws can be very costly.
1 	This 	report has 	been 	prepared 	by Dr. Alan 	Mathios.. 	It 	repres:nts
the 	views 	of 	the 	Bureaus 	of 	Competition, 	Consumer 	Protection. 	and
Economics., and does not necessarily represent 	the views of the Commission
or 	any 	individual 	Commissioner. 	The Commission 	has.. 	however. 	voted 	to
a uthorize its submission.
2 	WEFA. -An Evaluation of the FTCs Analysis of the Effects of RMA
Laws on 	Auto Markets- (January 	1987). 	The WEFA report was prepared at
the reQue~t of the National Automobile Dealers Association (NADA).
, 	State 	relevant 	market area 	laws 	in 	general 	give existing 	franchist.:d
automobile 	dealers a 	right 	to 	challenge 	the 	establishment 	or 	relocation 	by
their 	franchisor 	of 	other 	dealerships 	within 	a 	certain 	distance 	of 	their
dealership.

-
Section 	I of the 	report evaluates 	the 	theoreticil 	argu~nu that 	RMA
laws are desirable. 	Section II briefly discu1Ses the econometric model used
in 	tbe 	Staff 	Report. 	We 	then 	turn 	to 	\\'EFA's 	comments concerning 	the
Staff Report 	model 	WEFA's comments can be divided into three 	types: 1)
criticisms that apply to any econometric analysis, 2) comment! that can 	be
• empirically investig.ated to see if they cause ch~'nges in the conclusion of the
Staff Report, and 3) comments that cannot be addressed 	without additional
data, some of which can be obtained only from the auto dealers themselves.
We 	respond 	to these comments in 	Sections III, IV, and V respectively. 	We
present our conclusions in Section VI.
I. 	The Alleged Need for RMA Laws
WEFA discu~ the arguments put forth 	by proponents of 	RMA 	laws.
For 	example, 	VlEFA 	reports 	(at 	pp. 	8-12) 	that 	RMA 	laws 	are 	necessary
because 	without 	such 	laws: 	1) 	manufacturers 	will 	have 	an 	incentive
periodically 	to 	grant 	franchises 	in 	an 	area 	even 	when 	the 	market 	cannot
support 	an 	additional 	dealership, 	2) 	the 	value 	of 	goodwil! 	developed 	by
dealers 	is 	diminished 	by 	the 	establishment 	of 	new 	dealers., 	3) 	consumers'
.interests 	are 	not 	protected 	because 	divergence 	between 	the 	interests 	of
I
manufacturers 	and 	consumers 	causes 	a 	deterioration 	of 	dealership 	service
Quality, 	and 	4) 	the 	existing 	legal 	system 	does 	not 	provide 	adequate
protection 	of 	dealcn' 	rights. 	In 	addition, 	Vt'EFA 	claims 	that 	the .costs 	of
RMA 	laws 	are 	small 	in 	terms 	of 	dealerships 	that 	are 	denied. 	Below 	we
discuss each of these issues.·
• 	Vr'EFA 	provides 	no 	empirical 	evidence 	for 	any 	of 	the 	arguments
concerning the necessity for RMA la ws.

\\'EFA 	contellds 	(at 	p. 	II) 	that 	'w]hile 	the 	manufacturer 	cannot
maintain a healthy Ind producti .... e dealer organization if tbe dealerships, on
balance, 	are 	unprofitable, 	the 	manufacturer 	may 	benefit 	by 	granting 	new
francbises in a given market area even when the market cannot support an
additional 	dealersbip.- 	This 	argument 	ignores 	tbe 	incentives 	both
.manufacturers and dealers ha .... e to maintain a strong dealer network. 	It is
Dot in 	the 	interest of either prospecti .... e dealers or 	manufacturers to 	invest
in a market in which an additional dealer cannot prcfita bly exist. 	While a
-f1y-by-night- manufacturing company would engage in actions that decrease
the 	value 	of 	a 	franchise 	after 	a 	dealer-manufacturer 	agreement 	has 	been
reached, 	this 	is 	not 	the 	case 	for 	auto 	manufacturers. 	whos.e 	treatment 	of
dealers will be widely known and who must rely on dealers in the future. 	If
a 	manufacturer is too willing to experiment with additional dealerships, the
resulting dealership failures would make it more costly for- the manufacturer
to attract new dealers in the future. 	Therefore, it is in 	the manufacturer's
best 	interest 	to 	establish a strong dealer network. 	Even 	if a 	manufacturer
attempted 	to 	mislC3d 	potential 	dealers 	regarding 	the 	profit 	potential 	of 	a
new 	dealership, 	the 	prospective 	dealers 	have 	strong 	incentives 	to 	obtain
unbiased estimates of the potential sales of the mark.eL 	Therefore, whether
"'RMA laws exist or not., there is little reason 	to belie .... e 	that manufacturers
will have either the ability or incentive to open too many dealerships in any
markeL
WEFA 	is 	correct 	in 	its 	assertion 	that 	it 	is 	important 	for 	a 	dealer 	to
establish 	a 	favorable 	reputation 	(goodwill) with 	the 	public. 	However, it IS
not evident that the establishment of another dealership diminishes the \'all,;~
of 	the 	goodwill. 	For example, 	upon 	receiving 	high 	quality service 	from 	a
3

Chevrolet 	dealer 	in 	an 	uea 	wher~ 	there 	are 	two 	Chevrolet 	dealen,
purchasers 	are 	likely 	to 	recommend 	the 	particular 	dealer 	(rom 	whom 	tbey
purchased 	the car. 	In any event, WEFA correctly asserts (at p. 12) that ·an
individual dealer who establishes a good reputation 	will also establish value
for the manufacturer because of the linking of the dealership with a make of
vehicle (~.g.. Smith Chevrolet or Jones Dodge).· 	Because this is the case, it
is 'in 	the 	manufacturer's 	economic 	interest 	to 	keep 	this 	franchise 	healthy.
Actions by the manufacturer that would diminish the return to the individual
dealer would also hurt the manufacturer. 	Therefore, RMA la ws do not seem
necessary to prolect the value of a dealer's goodwill.
WEFA 	also 	claims 	(at 	p. 	7) 	that 	RMA 	laws 	are 	needed 	·beca use
manufacturer and consumer interests are not coincident and the unrestricted
ability 	of 	manufacturers 	to 	open 	new 	dealerships 	may 	result 	in 	a
deteriora tion 	in the quality and quantity of dealer servicc:...· 	This argument
is 	inconsistent 	with 	the 	workings 	of 	a 	competitive 	automobile 	markeL
Competition 	from dealers of other manufacturers will insure that consumers
receive the mix of service, quality, quantity and price they o;I'ant. 	In fact, it
is 	possible 	that tbe 	reduced level 	of competition in 	RMA states 	may cause
dealers 	to 	be 	less 	responsive 	to 	the 	desires 	of 	consumen, 	and 	provide 	a
'lower 	quality 	of service. 	In 	addition, 	manufacturers 	and 	consumers 	have
congruent interesu in the maintenance of dealer quality ~ince consumers are
unlikely 	to 	recommend 	the 	dealer 	to 	other 	consumers, 	or 	purchase 	the
manufacturer's 	product 	from 	the 	dealer 	in 	the 	future, 	unless 	they 	receive
adequate service from the dealer.'
£ 	Most 	car 	sales 	by 	dealers 	are 	probably 	concentrated 	in 	the
immediate local ar~. 	Therefore, ·word of mouth· communication concerning
the quality of the service may be an important determinant of sales.
4

WEFA argues that dealers' rights are not adequately protected in spite
of 	existing 	dealer-licensing 	laws 	and 	the 	dealer's 	ritht 	to 	sue 	the
manufacturer. 	For ex.ample, it is argued 	that use of the court system is so
costly to dealers that 	is docs not adequately protect their rights. 	But there
would seem to be nothing distinctive about automobile dealerships that makes
the protection of their rights under existing laws less adequate than those of
any 	other 	franchisee. 	Consequently, 	it 	is difficult 	to 	see why 	automobile
dealers 	in 	particular require special treatment under 	the law. 	In addition,
the 	establishment 	of 	an 	alternative 	to 	the 	court 	system 	that 	imposes 	low
costs 	on 	existing 	dealers 	can 	create 	incentives 	for 	those 	dealers 	to 	raise
spurious protests. 	This is particularly true if unsuccessful protests can delay
the opening of new dealership by several months. 6
Vr'EFA 	claims 	that 	existing 	data 	do 	not 	support 	the claim 	that 	RMA
laws 	will lead 	to higher prices by stopping the opening- of new dealerships.
The 	data 	'\\'EFA 	cites are 	from 	a 	National 	Automobile 	Dealers 	Association
survey 	which 	shows 	that 	only 	6 	percent 	of 	incumbent 	dealer 	protests
resulted 	in 	new 	dealerships 	being 	denied. 	However, 	the 	6 	percent 	figure
represents only the -tip of the iceberg- in terms of the impact of RMA laws
on the number of dealerships. 	First, RMA laws provide existing dealers with
>the opportunity to delay the 	opening of new ~ealerships by filing spurious
,
protests. 	Such deuys will reduce the number of dealershi~ at any point in
time. 	Second, 	the 	6% 	figure 	does 	not 	reflect 	dealenhips 	that 	the
manufacturer 	decided 	not 	to 	establish 	because 	of 	the 	expected 	costs 	and
6 	Vr'EFA 	cites 	a 	survey 	by 	the 	NADA 	which 	shows that 	only 6% 	of
protests 	are 	decided 	in 	favor 	of 	the 	protesting 	dealer. 	This 	is 	consisten t
with 	the 	notion 	tbat 	many 	of 	the 	protests 	are 	unfounded 	and 	are 	made
simply to delay tbe opening of new dealerships. See discussion infra.
5

delays a~iated with defending 	themselves against 	the protests 	of existing
dealers. f 	To correctly ascertain the effect of tbe RMA laws on the number
of 	new 	dealerships. one 	needs 	to 	know 	the 	number 	of 	dealers 	that 	would
have 	been 	established 	in 	the 	absence 	of 	RMA 	laws. 	The 	percentage 	of
protests decided in favor of the manufacturer gives the minimum number of
-dealerships 	not 	established 	because 	of 	RMA 	laws. 	The 	actual 	deterrence
effect on entry may exceed this minimum by a substantial amount.
In summa·ry, the theoretical arguments that RMA laws are desirable are
unfounded. 	RMA laws do not seem necessary for the proper functioning of
the 	automobile 	dealer 	market 	Manufacturers 	do 	not 	have 	incentives 	to
over-saturate 	a 	market 	with 	dealerships 	or 	to 	provide 	inferior 	Quality.
Moreover. 	RMA 	laws 	appear 	unnecessary 	for 	the. 	protection 	of 	dealers'
rights. 	Their rights are protected 	by existing laws and 	the right to sue 	in
the same fashion as those franchisees in any other indusir·y. 	Finally. rather
than 	providing evidence on 	the 	total 	number of dealerships 	that 	have 	not
been 	established 	because 	of 	RMA 	laws, 	WEFA 	has 	provided 	evidence
concerning 	the 	minimum 	number 	of 	dealerships 	that 	have 	not 	been
established.
Although 	the theoretical arguments on which WEF A relies are fIa wed,
'the actual effect of RMA la ws is still an empirical issue. 	It is to this more
important issue that we now turn our attention.
7 	In other words, the effect of the RMA laws cannot be judged by the
percentage of prot~ts that are decided in favor of the manufacturer. 	Those
manufacturers that Itteopt to establish new dealers in spite of the Rt.iA law
are 	those 	manufacturers 	who 	are 	relatively 	confic::nt 	they 	will 	\I.'in 	the
protesL 	Those manufacturers who feel 	that the RMA 	laws put them at 	too
much of a disadvantage 	may 	not even 	try to establish a dealership even 	i;
that 	would 	be 	in 	their 	(and 	the 	consumer's) 	best 	interest. 	Therefore, 	it 	is
not 	surprising 	that 	only 	6 	percent of the 	protests are 	decided 	in 	:avor 	of
the protesting dealer.
6

II. 	The Starr Report Model
To examine the effects of RMA laws on the retail price of automobiles,
the 	Staff 	Report 	developed 	a 	model 	of 	supply 	and 	demand 	for 	new
a utomobiles, 	using 	1913 	cross 	section 	da ta.' 	Unlike 	modeu 	developed 	by
• 	other~ 	the 	Staff 	Report 	model 	takes 	account 	of 	the 	possibility 	that
variations 	across 	areas 	of 	the 	country 	in 	the 	laws 	regulating 	auto 	dealers
may 	be 	in 	part 	due 	to 	different 	conditions 	in 	local 	auto 	retail 	markets.
Previous authors have assumed that the laws are established independently of
any influence by auto dealer~
In addition, the Staff Report model allows for different effects of RMA
laws depending on whether or not the region is experiencing an increase in
population. 	This 	latter 	consideration 	is 	extremely 	important. 	Laws 	that
restrict 	entry 	of' dealers 	into 	areas 	where 	there 	is 	uro 	or 	negative
population 	growth are likely to have no effect, because in tbose areas it is
unlikely 	that 	manufacturers 	would 	find 	it 	in 	their 	interest to 	start 	a 	new
.'
dealership 	anyway. 	Therefore, 	an 	econometric 	analysis 	that 	does 	not
distinguish between areas with negative and positive growth is likely to mask
some of the effect of RMA laws.
[he effect of RMA laws on the price of automobiles is estimated from
the 	price (supply) equation, within, a 	multivariate 	frame'work 	that 	accounts
for 	the 	effects 	of 	various 	factors 	on 	price. 	While 	taung 	account 	of
, 	This 	model 	is 	similar 	to 	the 	models 	of 	Smith 	and 	Eckard. 	See
Eckard, E.W.. Jr. ~e Effects of State 	Automobile :>ealer En:ry Regulation
on 	New 	Car 	Prices-, Economic 	Inquiry, 	Vol 	XXIV, 	No.2, (April 	1ge5), ~v·
223·42, 	and 	Smith, 	R.L.. 	-Franchise 	Regulation: 	An 	Economic 	Analysis 	of
State 	Restrictions 	on 	Automobile 	Distribution-, 	Journal 	of 	Law 	and
Economics, Vol XXV, (April 19&2), pp. 125·57.
7

population 	growth., the 	prices of automobiles 	in 	areas 	with 	RMA 	laws, are
compared 	to 	thos.e 	in 	areas 	wbere 	tbere 	is 	no 	la w. 	The 	Staff 	Report
concludes 	that 	in 	areas experiencing population 	growth, 	RMA 	laws 	ba ve 	a
large and statistically significant positive effect on the price of automobiles..
This 	is 	the 	conclusion 	that 	WEFA 	claims 	is 	invalid, 	becau.s.e 	according 	to
WEFA, tbe concl~ion results from a misspecification of the supply model.
III. 	WEF A Comllluts OD the Econometric Analysis
We address first 	WEFA's assertion that tbe Staff Report's econometric
analysis 	has 	specification 	errors. 	This 	is a 	relatively 	easy claim 	to 	make
with 	respect 	to any econometric analysis. 	It 	is 	more 	difficult 	to evaluate
the effect of any asserted specification error on the results of the analysis..
After carefully considering 	the misspecifications asserted 	by WEF A we are
unable 	to 	find 	any 	indication 	that tbe Staff 	Report omitted any 	variables
tbat 	would 	be 	expected 	to 	be 	correlated 	with 	RMA 	laws..9 	Absent 	a
relationship with RMA laws, the omission of such variables clearly does not
affect 	the 	validity 	of 	the 	analysis.. I
In 	addition, 	WEFA 	asserts 	that
relationsbips 	among 	included 	variables 	can 	lead 	to 	misleading 	or 	biased
results. 	We 	expl1in 	below 	why 	this 	is untrue. 	Finally, 	V{EFA 	mistakenly
~ equates 	the 	total 	explanatory 	power 	of 	an 	econometric 	model 	with 	the
significance of an individual variable.
\VEFA assem (at p. 4) that ·Viewed from the perspective of its ability
to 	explain 	regional 	variations 	in 	the 	dealer 	margin, 	the 	FTC 	model 	is 	a
failure. 	The 	model 	introduces 	108 [acton of which 	75% have an 	incorrect
9 	By correbtion 	we 	mean 	the regression 	coefficient of the 	RMA 13 w
In the regression of the excluded variable on all includec "·ariables.
8

(using FTC criterion) directional or I 	statistically insignificant effect on the
margin·. 	We 	believe 	this 	statement 	is 	misleading 	because 	it 	fails 	to
consider 	that 	if the 	factors 	in the 	model 	were 	not 	explaining 	varia tion 	in
dealer margins, us.ing common statistical tests. 95% of the factors 	would 	be
insignificant. 	In fact, 	a 	very 	simple statistic 	exists 	to 	determine 	whether
-the 	entire 	set 	of 	factors 	in 	a 	model, 	taken 	together, 	are 	statistically
meaningful. lO 	Each of tbe 9 equations in the Staff Report passes this test
at more than the conventional level of significance.ll
VlEFA also clainu (at p. ]4) tbat the Staff Report demand model -leads
to 	clearly 	incorrect 	estimated 	relationships· 	because 	-it 	introduces 	all 	the
separate 	relevant 	explanatory 	factors 	individually, 	[where] 	the 	various
explana tory 	factors 	are 	interrelated-. 	This 	is 	not 	true. 12 	In terrela ted
explanatory 	variables do 	not 	cause 	incorrect 	estimation (bias). 	In 	fact, 	if
the 	estimation 	yields 	significant 	coefficients 	they 	are 	perfectly 	valid 	and
10 	This 	statIStJC 	is 	called 	an 	F-statistic. 	The 	F-statistic 	is 	based 	on
how much variatioD in the variable of interest is explained by the factors in
the 	mc.del. 	For the 	two-stage 	least 	squares 	model 	(the 	model 	used 	in 	the
Staff Report) the relevant statistic is actually a chi-square statistic.
11 	We preseDt these statistics in Appendix C, where tbe results of the
dealer margin equation are presented.
12 	In 	fact, 	if 	a 	model 	has 	interrelated 	explanatory 	variables, 	the
omission 	of 	any of 	these 	variables 	from 	the 	analysis 	will 	bias 	the 	results.
To 	avoid 	misspecifiC3tion, 	it 	is 	essential 	that 	all 	variables 	that 	affect 	the
dependent 	variable Ind are interrelated 	with 	other explanatory variables be
included. See Pindyck.., R.. and RUbinfeld, D., Econometric Models & Econorr.:'c
Forecasts, McGraw-HilL. 198] (pp. ]28-130).
9

correct. 13 	An 	interrelationship among variables is an 	explanation 	only 	for
why individual coefficients may appear to be insignificanL
According 	to 	WEFA 	if 	a 	relevant 	variable H 	is 	excluded 	from 	the
analysis, tbe results concerning the included factors arc 	Dot valid. 	This 	is
not always true. 	The exclusion of relevant variables bia!>Cs the estimates of
the 	included 	factors 	only 	if 	the 	included 	factors 	are 	correlated 	with 	the
excluded 	relevant 	factor. u 	WEFA's 	criticism 	of 	tbe 	Staff 	Report 	is
unfounded 	because 	they 	provide 	no 	evidence 	of 	correlation 	between 	the
excluded 	and 	included 	factors. 	For example, WEFA 	suggests 	that 	s'pecific
regional 	costs of owning and operating an automobile dealership should 	be
included 	in 	the 	analysis. 	The 	issue 	is 	not 	whether 	these 	resional 	costs
affect 	price (they probably do), but 	rather whetber they differ for 	dealers
that operate under RMA 	laws and 	those that do not. 	If regional 	costs do
not 	differ 	in 	tbis, way, 	then 	the 	estimates obtained 	without 	including 	the
regional cost variables are unbiased.
WEFA suggests that if a model does not explain most of the varia tion
in the variable being studied, i.e.. have a high 	R', 	then 	the model 	has 	not
performed well lS 	For example, WEFA states (at p. 21) -the objective of the
IS 	See 	Kmenta. 	J.. 	Elements 	of 	Econometrics. 	Macmillan 	Pubiishing
,Co .. 	New 	York. 	New 	York. 	19'71. 	Kmenta 	states 	(p. 	388) 	-Let 	us 	now
examine 	the 	connection 	between 	tbe 	degree 	of 	multicollinear::y 	and 	the
properties 	of 	the 	least 	squares 	estimators 	of 	the 	regression 	coefficients.
Under the basic assumptions of the classical nQrmal linear regression model,
the 	least 	squares 	estimators 	of 	the 	regression 	coefficients 	have, all 	the
desirable properties.-
H 	By relevant. 	we 	mean 	that the variable is 	important 	in explaining
the dealer price or margin.
See Pindyck. R .. and RUbinfeld, D .. pp. 128·130.
16 	R2 indiC3tes 	tbe 	percentage of variation 	in 	the dependent 	variable
explained by the model.
10

FTC supply 	model was to explain 	regional variations 	in the dealer margin
and 	to determine whether introducing RMA la W$ increased this margin. 	If
essentially all the explanatory power in the average unit retail price equation
comes from the unit dealer cost variable. the FTC model is not doing a good
job at explaining variation in the dealer margin-. 	If the Question is whether
• RMA laws increase the retail price of automobiles, a statistically significant
positive relationship between retail price and RMA laws is entirely sufficient
to conclude that [bey do. l1 	The percentage of the variation in retail price
explained by the Staff Report model is not a relevant consideration.ll
The 	\VEFA 	analysis 	next 	turns 	to 	more 	substantial 	matters 	of
misspecification. 	The major misspecification cited by WEFA can be analyzed
by simple respecifications of the model. 	This is tiiscussed below.
IV. 	Empirical IDYutigatioD of WEFA Comments
The WEFA report claims that. rather than finding that RMA laws result
in higher prices in areas that are experiencing large percentage increases in
population. 	the Staff 	Report has 	really 	found 	that either. 	1) market 	areas
with 	large 	absolute 	population 	increases 	have 	higher 	retail 	motor 	vehicle
prices. or 2) market areas with old RMA laws and large absolute population
"increases have higher retail motor vehicle prices. 	In particular Vr'EF A notes
11 	This conclusion assumes tha t there are no omitted facto,s 	that are
correlated 	with 	the 	laws. 	There 	are 	many 	examples 	of 	empirical 	models
where there is a low R 2 and very reliable effects on a specific factor. 	For
example. 	in 	examining 	earnings 	across 	individuals, 	those 	who 	have 	more
experience 	earn 	more 	than 	those 	with 	less 	experience. 	The 	typical 	R 2 in
such 	models 	is 	approximately 	.2 	yet 	the 	factor 	'experience- is 	statistically
significant.
18 	The 	measure 	of 	goodness-of·fit 	used 	in 	the Staff 	Report 	modd 	is
actually 	the 	square 	of 	the 	correlation 	coefficient 	between 	the 	actual 	and
predicted value of the dependent variable.
1 I

(at 	p. 	4) 	that ·since 	only 	large 	rapidly 	growin& 	cities 	have 	large 	absolute
population 	increases. the FTC may merely 	have documented 	that 	wage 	and
other 	dealer 	costs 	are 	higher 	in 	major 	metropolitan 	areas 	than 	in 	smaller
urban areas aDd rural areas.·
The Staff Report included the percentage change in population (whether
positive or negative) as a factor to explain retail prices. 	WEFA asserts that
the absolute change in population for areas with growing population should
have 	been 	used 	a..s a 	factor. 	since 	a 	large 	growing 	city 	can 	have 	a 	small
percentage increue in population but a large absolute increase. 	To account
for 	this 	we 	have re-estimated 	the 	identical 	price 	(supply) equation, except
that this time we have included a variable that captures whether an area has
pOSItive absolute population growth and. if so, how large an absolute growth.
Our results confirm the conclusion drawn from the original equation, namely
that 	RMA 	laws 	in 	areas 	with 	growing 	populations 	(Whether 	absolute 	or
percentage) have a positive effect on motor vehicle prices. 1Q
We 	have 	also 	re-estimated 	the 	identical specification, except 	that 	this
time we use the percentage increase in population only if positive. and zero
otherwise. 	This will 	allow 	us 	to 	compare 	prices 	in 	areas with 	population
growth and RMA laws to prices in areas with population growth but without
IRMA 	laws. 	The results 	using this 	variable confirm 	the 	conclusion 	that 	in
growing areas. RMA laws have a positive and significant effect on the price
of a utomobiles.~
lQ 	The ne .... and 	original regression 	results are 	presented 	in Appendix
A, along with a brief summary of the findings regarding the law variables.
:0 	The 	regression 	results 	for 	the 	originaJ 	and 	new 	specifications 	(l:.-e
given 	in Appendix B along with a 	brief summary of findings regarding 	the
la w varia bles.
12

The WEFA study suggests that it is more appropriate to explain dealer
margins than to explain retail prices. 	WEFA claims that by including dealer
cost as an explanalory variable, the R S of the -equation is bither and offers
a misleading indicator of performance of the model. 	While we do not agree
with 	WEFA's 	c1aim 21 , 	in 	the 	interest of completeness 	we 	bave 	re-estimated
• 	the equation with dealer margin as tbe dependent variable to be explained.
The results confirm our conclusion 	that in growing areas RMI\ laws have a
positive effect on automobile prices.:!
V. 	WI:fA Commut' that Require Additional Data
We 	agree 	with 	WEFA 	that 	it 	is 	preferable 	to 	include 	dealer 	cost
variables that are associated with the specific market area (S~1SA or county
level) 	than 	to 	use 	statewide 	averages. 	For 	example, 	WE FA 	suggests 	that
wages should 	be collected 	for each 	market area, but, as vtEFA 	notes, such
data 	may 	not 	exist. 	In 	lieu 	of specific 	wage 	data, 	V{EFA 	suggests 	using
wholesale-retail trade and service-sector wage rates, which can be calculated
for 	all 	the market areas. 	It is not clear whether these 	variables would 	be
superior 	to 	statewide averages of 	wages specific to 	the 	auto 	dealer 	sector.
This would depend on how similar auto-dealer wages are to tbe entire retail·
(
21 	We have already dismissed the role of R% in evaluating the impact
of RMA 	laws on the 	price of motor vehicles. 	In addition, the econometric
model 	v,'as 	originally 	specified 	as 	a 	demand 	and 	supply 	equation 	for
automobiles.. 	Consumers do not 	care about 	margins, 	they care about price.
It is appropriate to estimate the mode! with price as the dependent variable
and 	include 	determinants 	of 	price 	8~ explanatory 	variable!.. 	In 	fact, 	by
including 	the 	cost 	of 	the 	automobile 	to 	the 	dealer 	as 	an 	explanatory
variable, 	the 	Staff 	Report 	is 	being 	less 	restrictive 	since 	is 	not 	restricting
the 	coefficient 	on 	this 	factor 	to 	be 	equal 	to 	I, 	as 	is 	the 	case 	if 	dealer
margin is used as the dependent variable.
%1 	The results of these regressions arc given in Appendix C along with
a brief summary of the findings regarding the law variables.
13

-._---
wholesale 	trade 	wages 	versus 	how 	similar 	local 	auto 	dealer 	WI ges 	are 	to
statewide auto dealer wages.
WEFA 	also 	suggests 	that 	it 	would 	be 	useful 	to 	obtain 	data 	for
dealerships over time and compare the prices of automobiles before the RMA
laws were passed to the prices afterward1. 	They claim that because many of
the 	potential 	omitted 	variables 	would 	not 	change over 	tim<; 	by 	comparing
pre- 	and 	post-law 	prices 	we 	would 	avoid 	many 	problems.. 	While 	we 	agree
that 	this 	would 	be 	a 	useful 	additional 	exercise, 	the 	lack 	of 	such 	a
comparison 	does 	not 	imply 	that 	the 	cross-sectional 	analysis 	in 	the 	Staff
Report 	is 	invalid.. 	In 	fact. 	in 	the 	time-series, 	cross-sectional 	analysis
suggested 	by 	\\'EFA. 	the 	number 	of 	cross 	section 	observations 	(different
deaiers) will far exceed the number of time periods (the number of years of
data 	for 	each dealer). 	Consequently. the 	bulk 	of 	the 	econometric 	analysis
would still compare prices across dealers, rather than prices' across time.
In 	summary. 	WEFA 	suggests 	that 	to 	·fix· 	the 	Staff 	Report 	model.
regional differences in dealer ownership and operating cost must be properly
and _completely 	incorporated 	into 	the 	model. 	Additionally. \\'EFA 	suggests
that 	regional 	differences in 	dealership size, sales mix 	within 	makes. dealer
supplied options., and regional differences in legislation must be included. 	As
'discussed above. because \\'EFA provides no evidence that these variables are
systematically related to the existence of RMA laws. it cannot be concluded
that their omission from 	the Staff Report model 	will affect the 	validity of
the conclusions about the effects of RNtA law~.
14

VI. 	CODcluloa
Although WEFA 	has 	claimed 	that the Staff Report 	may 	be subject 	to
specification error, WEFA provided no examples of misspecification that are
likely 	to 	affect 	tbe 	report's conclusion. 	WEFA 	cited 	a 	few 	variables 	tbat
are not included in the analysis but Vr"EFA provid=s 	no evidence that these
variables differ systematically between regions that are subject to RMA laws
and 	regions that are not. a necessary condition for obtaining biased 	results.
WEFA's speculation 	that 	the Staff 	Report 	has 	demonstrated 	only 	tl.3t
rapidly growing areas have higher automobile prices 	has 	been shown 	to be
unfounded. 	When 	controlling for 	the absolute 	growth 	of an 	area. 	we still
find that RMA laws in areas with growing populations have a positive effect
on" motor 	vehicle 	prices. 	Other specifications of 	the 	growth 	variable 	also
confirm the effects of Rt.iA 	laws. 	Finally, examina tion of dealer margins
instead of prices does not change the report's concl usion. - -
15

Appndlx A
In this appendix we report the original and DCW estimates of the supply
equation. 	The following change was made in the equation. 	The growth rate
varia ble 	(GR) 	was 	originally defined 	as 	the 	logarithm 	of 	the 	ratio 	of 	the
observation year population to that in 1970 (pcrcentage chaDge in population
since 	1970). 	The new growth variable is defined as the absolute change in
population 	between 1970 and 	1978. if positive. aDd zero otherwise." 	All of
the 	other 	variables 	are 	defined 	as 	they 	were 	in 	the 	original 	report.
Obviously. 	this 	change 	will 	affect 	the 	coefficicnt 	on 	the 	growth 	variable.
Table 	reports 	the 	old 	and 	new 	coefficient 	estimates 	for 	the 	supply
eQua tion. 	A Comparison of the old and new results reveals that. while 	the
coefficients 	involving the 	RMA 	law variables (ROl.p. lAGRO. IAGR 1. and
IAGRlSQ) have changed slightly. the new estimate of the coefficient on the
interaction 	of 	the; 	law 	variable 	and 	absolute 	population- growth. 	IAGRI. 	is
larger 	and 	more 	significant for 	every car 	model. 	In 	Table 	1. coefficients
with 	t-values 	(which 	appear 	in 	parentheses 	under 	the 	coefficient) 	greater
than 	or 	equal 	(in 	absolute 	value) 	to 	1.96 are ,significant at 	the 	95% 	level.
Table 2 presents the old and new estimates of the impact of RMA la W$
on 	the 	price of each model type.:3 	It is clear from Table 2 that 	for 	both
, the old and new equations. RMA laws in areas of growing populations have a
positive impact on price. 	In addition, for 7 of the 9 model types. the 	new
equation yields larger positive price effects than did the old.
:: 	To 	be 	consistent 	with 	the 	original 	specification 	we 	have
transformed the absolute growth variable. 	The transformation is idcntical to
the transformation of the AGR variable in the Staff Report (p. 62).
:~ 	This 	estimated 	effect 	involves 	a 	combin:ltion 	of 	some 	of 	the
coefficient estimates of the supply equation.
16

Appendix A - Table I
The Estimated Regression Coefficients for the Retail Stlpply Equation
for the Nine Chevy Auto Body-Types for 1971
(Dependent Variable • Log of Retail Pri~)
Variable 	Body lYRe
Regular 	Malibu 	Cama fO
Qli1 	~ 	Qlit 	&:::. 	QM 	~
C 	.7875 	.68JO 	.5426 	.45 J2 	.2~:6 	.1447
RNEW 	-.0028 	-.0027 	-.0032 	-.OO:!8 	.0032 	.0037
(-.94) 	(-0.83 ) 	(-1.l6) 	(-0.90) 	(J.10) 	( 1.07)
ROLD 	-. J0 J3 	-.1253 	-.0941 	-.1317 	.0395 	-.0218
(-1.68) 	(-2.JO) 	(-1.70) 	(-2.32) 	(.66) 	(-0.34)
lAGRO 	.0103 	.0124 	.0095 	.0130 	-.0067 	-.0004
(1.49) 	(1. 79) 	( 1.50) 	( 1.97) 	(-.97) 	(-0.57)
IAGRI 	2.0869 	3.1406 	1.6992 . 	2.7910 	2.8680 	4.0269
(4.72) 	(5.70) 	(4.17) 	(5.46) 	(6.65) 	(7.08)
IAGR J SQ -5.348 	-9.5244 	-4.530 	-7.5770 ..
-9.169 	-13.40J
(-1.9J) 	(-5.J4) 	(-1.75) 	(-4.36) 	(-3.~2) 	(-7.10)
PW 	.9077 	.9 J62 	.9306 	.938 J 	.9926 	.9949
(69.83) 	(67.25) 	(77.80) 	(70.06) 	(99.74) 	(E8.82)
WAGE 	.0 J2 I 	.OJ 58 	.0208 	.0235 	-.0 I J8 	-.00~8
( 1.32) 	( 1.76) 	(2.4 3) 	(2.73) 	(-1.28) 	(-0.5 I )
ADV 	-.0043 	-.0055 	-.0038 	-.0054 	-.00J4 	-.0035
(-3.6J) 	(-5.00) 	(-3.43) 	(-5.04) 	(-1.J4) 	(-3.01)
,DENS 	-.0039 	-.00:!8 	-.00:2 	-.0015 	.0014 	.00:2
(-5.46) 	(-4. J8) 	(-3.38) 	(-2.26) 	(1.95) 	(3.05)
GR 	-.0031 	-.1755 	.0053 	-.1993 	.0150 	-.174 J
(-.40) 	(-2.02) 	(.73 ) 	(-2.41) 	(1.96) 	(-1.97)
,
FORCE 	-.0135 	-.0162 	-.0121 	-.0170 	-.OOi 1 	-.0106
(-2.53) 	(-2.70) 	(-2.39) 	(-:.90) 	(-1.41) 	(-I.ii)
CANCEL 	.01 i4 	.0226 	.01 i2 	.0241 	.0074 	.0137
(3.34 ) 	(3.8 J) 	(3.52 ) 	(4.19) 	(1.48) 	(2.29)
LD 	-.0131 	-.0107 	-.0160 	-.0146 	-.0116 	-.0090
(·2.47) 	(-1.90) 	(-3.14) 	(-2.65) 	(-:.::) 	(-1.54)
17

Appendix A - Table 1 (Continued)
The Estimated Regression Coefficients for the Retail Supply Equation
for the Nine Chevy Auto Body-Types for 1978
(Dependent Variable • Log of Retail Price)
Varjat-Ie 	Body Type
~ova 	Monte Carlo 	Monza
Ql.d. 	~ 	Ql.d. 	~ 	Ql.d. 	fu:=:.
C 	.7723 	.6915 	04099 	.3389 	.6682 	.5300
RNEW 	-.0006 	-.0006 	-.00:3 	-.0022 	-.0045 	-.0041
(-.20) 	(-.21 ) 	(-.82) 	(-.7:!) 	( 1.28) 	(-1.12)
ROLD 	.0155 	-.0260 	-.0990 	-.1184 	-.0231 	-.0951
(0.30) 	(-0.51 ) 	(-1.79) 	(-2.21 ) 	(-.32 ) 	(-1.42)
lAGRO 	-.0024 	.0021 	.0100 	.0120 	.0008 	.0086
(-040) 	(0.35) 	( 1.58) 	(1.91 ) 	(.09) 	(1.10)
IAGRI 	1.5946 	2.1550 	1.7007 	2.3502 	1.3734 	2.6807
(4.06) 	(4.60) 	(4.02) 	(4.59) 	(2.70) 	(4.64)
IAGR ISQ :'5.024 	- 7.4424 	-4.621 	-7.7015 - - -.9754 	-8.0685
( -2.03) 	(-4.67) 	(-1.73) 	(-4.52) 	(-.31) 	(-4.14)
PW 	.9067 	.9112 	.9495 	.9540 	.9235 	.9275
(90.34) 	(87.77) 	(81.92) 	(79.95) 	(104.7) 	(104.34 )
WAGE 	.0151 	.0207 	.0172 	.0213 	.0103 	.0243
(1.72) 	(2.57) 	( 1.96) 	(2.57) 	(.95) 	(2.50)
'ADV 	-.0021 	-.0035 	-.0027 	-.0037 	-.0015 	-.0044
( -1.96) 	(-3.47) 	(-2.44) 	(-3.63) 	(-1.04) 	(-3.58)
DENS 	-.0007 	-.0003 	-.0021 	-.0014 ~
-.0016 	-.0003
(-1.10) 	(-<l.56) 	(-3.05) 	(-2.25) 	(-1.84) 	(-0.4 1)
GR 	.0108 	-.0768 	.0017 	-.0908 	.0201 	-.1485
(1.54 ) 	(-1.02) 	(.23) 	(-1.13) 	(2.:2) 	(-1.64)
FORCE 	.0031 	.0013 	-.0162 	-.0172 	-.0062 	-.0075
(0.63) 	(0.24 ) 	(-3.22) 	(-3.14) 	(-1.00) 	(-1.17)
CANCEL 	.0036 	.0071 	.0179 	.0206 	.0128 	.0183
(.76) 	( 1.35) 	(3.68) 	(3.83) 	(2.09) 	(2.88)
LD 	-.0087 	-.0076 	-.0121 	-.0103 	-.0168 	-.0118
(-1.82) 	(-1.54) 	(-2.37) 	(-1.96) 	(-2.77) 	(-2.0:)
I!;

Appendix A - Table 1 (Continued)
The Estimated Regression Coefficients for the Retail Supply Equation
for the Nine Chevy Auto Body-Types for 197&
(Dependent Variable-Log of Retail Price)
Variable 	Body Tvoe
Cheven, 	SPQrt v2D 	CQrvette
Qli1 	~ 	Qld. 	~ 	Ql.d. 	New
C 	.5561 	.4336 	.2245 	.1123 	AI31 	.2340
RNEW 	-.0025 	-.0027 	-.0077 	-.0080 	.0033 	.0053
(-.81 ) 	(-0.77) 	(-1.61) 	(-1.58 ) 	(.36) 	(0.57)
ROLD 	-.0672 	-.1037 	-.0280 	-.0831 	.3033 	.3303
(-1.03) 	(-1.61 ) 	(-.24) 	(-0.79) 	( 1.69) 	( 1.97)
lAGRO 	.0053 	.0089 	.0012 	.0077 	-.0402 	-.0447
(.70) 	(1.18) 	(.09) 	(0.62) 	(-1.95) 	(-2.30)
IAGRI 	1.9861 	3.2553 	1.7323 . 	2.1 ~92 	3.1587 	6.4510
(4.33) 	(5.65) 	(3.45) 	(3.59) 	(:.36) 	(4.24 )
IAGR1SQ' -4.4453 	-9.9334 	-3.69i2 	-7.4675 - - -1.7200 	-15.84
(-1.54) 	( -5.21) 	(-1.31 ) 	(-3.69) 	(-.21 ) 	(-3.08)
PW 	.9289 	.9368 	.9649 	.9688 	.9743 	.9806
(95.60) 	(8&.95) 	(89.67) 	(89.24) 	(39.72) 	(39.62)
WAGE 	.0198 	.0268 	.0222 	.0336 	.0044 	.0186
( 1.99) 	(2.81 ) 	( 1.39) 	(2.33) 	(0.15) 	(0.73)
. ADV 	-.0038 	-.0057 	-.0056 	-.0070 	.OOC2 	-.0035
(-2.99) 	(-4.81 ) 	(-2.76) 	(-3.96) 	(0.06) 	(-1.07)
PENS 	-.0005 	.0008 	-.0018 	-.0006 	-.0087 	-.00~6
(-.66) 	(1.09) 	(-1.72) 	(-0.61 ) 	(-3.82) 	(-2.33)
GR 	-.0022 	-.1941 	-.0060 	-.0602 	-.0169 	-.5242
(-.27) 	(-2.13) 	( -.46) 	(-0.68) 	(-.70) 	(-2.: 1)
FORCE 	-.0096 	-.0121 	-.0120 	-.0097 	-.0031 	-.0041
(-1.71) 	(-1.93) 	(-1.60) 	(-1.28) 	(-.19) 	(-0.25)
CANCEL 	.0186 	.0246 	.02:2 	.0234 	.01 :4 	.0217
(3.3&) 	(3.92) 	(3.25) 	(3.34) 	(.79) 	( 1.3:)
LO 	-.0212 	-.0174 	-.0050 	-.0030 	.0::: 	.0344
(-3.93) 	(-3.10) 	( -.67) 	( -0.40) 	(1.37) 	(:.1 ~)
19

Appendix A • Ta ble 2
The Impact of the RMA La W$ on the Price of Chevrolet Cars in 1978
Body Type 	Average Percentage Price Change
For Areas with Positive POQulation Growth
Regular 	5.90·· 	7.68 --
Malibu l 	5.02·· 	8.35··
Camaro 	8.43·· 	11.80··
Nova 	4.18· 	5.09-
Mon te Carlo 	5.30·· 	6.03··
Mooza 	8.81·· 	7.08-·
Chevette 	5.87·· 	7.58
. Sportvao 	3.68 	3.28
Corvette 	16.82·- 	18.75
1 The esti~a te of tbe price effect for the Malibu 10 the Staff Report should
. ha ve been reported as 5.02 rather than 14.31.
• sigoifi:antly above zero at the 95% level.
•• significantly above zero at the 99% level.
:0

."
Appndlx B
In 	this 	appendix 	we 	report 	the 	original 	and 	new 	estimates 	of 	another
specification of the supply equation. 	The following change .-as made in the
specifica tion. 	The 	growth 	rate 	variable, 	GR, 	was 	changed 	from 	the
percentage 	growth 	rate 	to 	the 	growth 	rate 	if 	positive. otherwise 	zero. 	All
of the other variables are defined as they were in the origirul report. 	Table
J reports the old and 	new coefficient estimates for 	the supply eQua tion. 	It
is 	clear 	from 	the 	results 	below 	that 	the 	eQua tion 	is 	not 	s.ensitive 	to 	this
change 	in 	the 	growth 	rate 	varia ble. 	Like 	the 	original 	equation, 	the
interaction of the law ""ariable and areas with population growth. lAGRI. is
positive 	and 	significant 	in 	all 	9 	equations. 	In. 	the 	following 	tables.
coefficients with t-values of greater than or equal (in absolute value) to 1.96
are 	significant 	at, the 	95% 	level. 	The 	t-values 	are 	reported 	under 	the
coefficients and appear in parentheses.
Table 2 pre~nts the old and new estimates of the impact of R?>1.A la ws
on 	the 	price of each model 	type." 	It is clear from Table 2 that 	for 	both
the old and new equations, RMA laws in areas of growing popul:<tions have a
positive impact Ot price.
" 	This 	estimated 	effect 	involves 	a 	combination 	of 	so~: 	of 	the
coefficient estimates of the supply equation.
21

Appendix B - Table J
The Estimated Regression Coefficients for the Retail Supply Equation
(or the Nine Chevy Auto Body-Types for 197&
(Dependent Variable • log o( Retail Price)
Variable 	Bodv Type
R,gular 	Malibu 	Ca rna rQ
QM 	~ 	QM. 	~ 	Qli1 	~
C 	.7875 	.7803 	.5426 	.5327 	..2.226 	.2119
RNEW 	-.0028 	-.0029 	-.0032 	-.0032 	.0032 	.0032
(-.94) 	( -.98) 	(-1.16) 	(-1.21 ) 	(J.l 0) 	(1.08)
ROLD 	-.1013 	-.1091 	-.0941 	-.1060 	.0395 	.0196
(-1.68) 	(-1.84) 	(-1.70) 	(-1.93) 	(.66) 	(.33)
lAGRO 	.0103 	.0 112 	.0095 	.0109 	-.0067 	-.0044
(1.49) 	(1.65) 	( 1.50) 	( 1.73) 	(-.97) 	(-.64)
IAGRI 	2.0869 	2.073 	1.6992 	1.6939 	2.8680 	2.888
(4.72) 	( 4.69) 	(4.17) 	(4.14) 	(6.65) 	(6.69)
IAGR 1SQ-5.348 	-5.39 	-4.530 	-4.619 	---9.169 	-9.315
(-1.91 ) 	(-1.93) 	(-1.75) 	(-1,78) 	(-3.42) 	(-3.47)
PW 	.9077 	.908 	.9306 	.9314 	.9926 	.9924
(69.E3) 	(69.78) 	(77.80) 	(77.43) 	(99.74) 	(99.54)
WAGE 	.0121 	.012 	.0208 	.0213 	-.0118 	-.0102
(1.32) 	(1.33) 	(2,43 ) 	(2.49) 	(-1.28 ) 	(-1.11)
ADV 	-.oo·n 	-.0043 	-.0038 	-.0389 	-.0014 	-.0017
(-3.61) 	(-3.66) 	(-3,43) 	(-3.54) 	(-1.14) 	(-1.42)
DENS 	-.0039 	.0039 	-.0022 	-.0022 	.0014 	.0013	, (-5.46) 	(-5.41) 	(-3.38) 	(-3.34) 	(1.95) 	( 1.86)
GR 	-.0031 	-.0024 	.0053 	.0050 	.0150 	.0095
(-AO) 	( -.30) 	(.73) 	(.66) 	(1.96 ) 	( 1.19)
FORCE 	-.0135 	-.0138 	-.0121 	-.0125 	-.0071 	-.0075
(-2.53) 	(-2.58) 	(-2.39) 	(-2,45) 	(-1.41 ) 	(-1,49)
CANCEL 	.0 j '74 	.0177 	.01'72 	.0177 	.0074 	.0080
(3.34) 	(3.39) 	(3.52) 	(3.59) 	(I.4S) 	(1.60)
LD 	-.0131 	-.0132 	-.0160 	-.0161 	-.0116 	-.0116
(-2.47) 	(-2.48) 	(-3.14) 	(-3.1S} 	(-2.::) 	(-:.::n
............

--_.. _--- 	_..

Appendix B • Table 2
The Impact of tbe RMA Laws on the Price of Chevrolet Cars in 1978
Body TyPe 	Average Percentage Priu Change
for Areas wjeh Positive Population Growth
Regular 	5.90" 	5.79·'
Malibu l 	5.02'· 	«.69"
Camaro 	8.43·' 	8.44"
Nova 	4.18 • 	...06'
Monte Carlo 	5.30·' 	5.28·'
Monza 	8.81 •• 	8...64 ••
Chevette 	5.87" 	5.81·'
Sportvan 	3.68 	3.89
Corvette 	16.82" 	17.28 •
1 The estimate of tbe price effect for the Malibu in the Suff Report should·
ha ve been reponed as 5.02 rather than J4.31.
• significantly above z.ero at the 95% level.
•• significantly above z.ero at the 99% level.

Appndlx C
In this appendix we report tbe estimates of the dealer margin equation.
As opposed to the original specifications. the following table uses the dealer
margin 	as 	the dependent 	variable. 	Since the dealer's 	unit cost (cost 	of 	the
car to the dealer) is part of the dependent variable, it is no longer used as
either a 	regres.sor or as an instrumental variable. 	The growth rate variable
as specified 	in appendix B (rather than the absolute 	growth varia ble or the
original 	growth 	variable) 	was 	used 	in 	the 	analysis. 	All 	of 	the 	other
explanatory 	variables are defined 	as they were in 	the original report. 	The
dependent 	variable 	is simply defined 	as 	the difference 	of the 	log of retail
price 	and 	the 	log of 	dealer's cost. 	This 	represents 	the 	percent 	markup 	to
the dealer. 	Obviously, the estimated coefficient vector will change since we
are nowexplaini~g dealer margins. 	The results in Tabl~-I show that in 8 of
the 9 model types the interaction of the law and population growth variable,
IAGRI. is still 	positive and 	significant. 	In 	the 	tables. 	coefficients 	with 	t-
values of greater than or equal (in absolute value) to 	1.96 are significant at
the 	95% level. 	The 	t-values are 	given 	under 	the 	coefficient and 	appear 	In
parentheses.
In response to WEFA's claim that the Staff Report model is a failure in
explaining 	dealer 	margins. 	we 	present 	the 	chi-square 	statistic, 	which 	is 	a
test 	for 	whether 	all 	the 	factors. 	when 	taken 	together, 	are 	stalistically
significant. 	The critical chi-square value for the 99% significance Icve! in a
26

model 	with '12 facton 	is 	26.22. 	The 	re,ult, 	In 	Table 	1 show 	that 	all 	the
models far exceed this critical value.u
Table 	2 	reports 	the 	total 	impact 	of 	RMA 	laws 	in 	areas 	of 	growing
population 	on 	the 	dealer 	margin 	for 	each 	model 	type.~ 	We 	obtain 	a
statistical1y 	significant 	positive 	impact 	on 	dealer 	margins 	for 	7 	of 	the 	9
model types.
:6 	In 	this 	appendix 	we 	use 	the 	chi-square 	statistic 	to 	test 	the
significance of the model. 	The measure of goodness-of -fit used in the Staff
,Report 	was 	the SQuare of the correlation 	between 	the 	actual 	and 	predicted
yalue of the dependent variable. 	However. even for a 	well specified model.
we 	would 	not 	always 	expect 	a 	high 	correlation 	between 	the 	actual 	dealer
margin and the predicted dealer margin especially when the prediction comes
only 	from 	the supply 	equation 	and 	ignores 	the 	demand side 	of 	the 	model.
For example. 	in 	the dealer 	margin 	equation 	presented in this appendix. the
correlation between predicted and actual dealer margin for the 9 models are
small 	(they 	range 	from 	.05 	to .17 	_. 	the 	square 	of 	these 	numbers 	will 	be
close 	to 	zero) 	while 	the 	chi-squae statistics 	reveal 	that 	the 	factors 	in 	the
model are significant at more that the 99% level of significance.
:6 	This 	estim:lted 	effect 	involves 	the 	combination 	of 	sam: 	of 	the
estimated coefficients of the supply eGuation.

Appendix C - Ta ble 1
The Estimated Regression Coefficients for the Retail Supply Equation
(or the Nine Chevy Auto Body-Types for 1978
(Dependent Variable-Dealer Margin)
Variable 	Body Type
Monte
R egular 	D.r.1Q 	Maljbu 	CamarQ 	~
C 	-.0450 	-.0401 	-.0854 	.1379 	-.1003
RNEW 	-.0053 	-.0025 	-.0050 	.0028 	-.0035
(-1.75) 	(-0.88) 	(-1.73) 	(0.96) 	(-1.25)
ROLD 	-.1795 	-.1311 	-.1715 	.0133 	-.0955
(-3.01) 	(-2.40) 	(-3.05) 	(0.22) 	(-1.85)
lAGRO 	.0195 	.0139 	.0183 	-.0037 	.0106
(2.85) 	(2.21 ) 	(2.84) 	(-0.54) 	( 1.78)
IAGRI 	1.7217 	1.5557 	1.4993 	2.9117 	1.1881
(3.73) 	(3.59) 	(3.43) 	(6.69) 	(2.87)
IAGRISQ 	-5.7698 	-4.9421 	-4.7864 	-9.6570 	-4.8222
(-1.98) 	(-1.82) 	(-1.73) 	( -3.59) 	(-1.85)
WAGE 	.0192 	.0204 	.028\ 	-.0087 	.0317
(2.04) 	(2.31 ) 	(3.11 ) 	(-0.96) 	(3.59)
ADY 	-.0042 	-.0027 	-.0041 	-.0018 	-.0030
(-3.42) 	(-2.42) 	(-3.55) 	(-1.55) 	(-2.70)
DENS 	-.0033 	-.0016 	-.0020 	.0014 	-.0001
(-4.47) 	( -2.34) 	(-2.88) 	(2.04) 	(-<l.18)
GR 	-.0041 	.0013 	.0034 	.0098 	.0114
(-0.49) 	(0.17) 	(0.42) 	( 1.22) 	(1.48)
'FORCE 	-.OJ i7 	-.0188 	-.0186 	-.OOS 1 	-.0064
(-3.22) 	(-3.71) 	(-3.57) 	(-1.61) 	(-1.28)
CANCEL 	.0207 	.0188 	.0230 	.0090 	.0118
(3.87) 	(3.82) 	(4.53) 	( 1.82) 	(2.44)
LD 	-.0197 	-.0161 	-.0238 	-.0127 	-.0184
(-3.65) 	(-3.15) 	(-4.64) 	( -:.49) 	(-3.n)
Chi
Square 	106.64 	105.65 	138.83 	209.09 	123.85
:8

, 	.
Appendix C - Table 1 (Continued)
The Estimated RegressioD Coefficients for the Retail Supply Equation
for the Nine Chevy Auto Body-Types for 1971
(Dependent Variable-Dealer MArgin)
VjHjable 	Bodv Type
Moon 	Chevette 	SpQrtvan 	Corvette
C 	-.0422 	-.1281 	-.1444 	.1645
RNEW 	-.0080 	-.0062 	-.0086 	.0032
(-2.18) 	(-1.95) 	(-1.80) 	(0.36)
ROLD 	-.1443 	-.1596 	-.1096 	.2982
(-1.95) 	(-2.44) 	(-1.02) 	(1.71)
lAGRO 	.01497 	.0162 	.0109 	-.0395
( 1.76) 	(2.16) 	(0.87) 	(-1.97)
IAGRI 	.8204 	1.7638 	.. 1.6019 	3.1715
( 1.54) 	(3.67) 	(3.20) 	(2...38)
IAGR1SQ 	.7973 	-4.1036 	-3.9513 	-1.9699
(0.24) 	(-1.36) 	(-lAO) 	(~.24)
WAGE 	.0027 	.0360 	.0321 	.0066
(2.05) 	(3.61 ) 	(2.05) 	(0.24)
ADY 	-.0027 	-.0054 	-.0067 	.0001
(- 1.8 I) 	(-4.25) 	(-3.41) 	(0.03)
DENS 	-.0014 	-.0004 	-.0014 	-.OOD
(-1.50) 	(-0.54 ) 	(-1.28) 	(-3.71 )
~GR 	.0181 	-.0055 	-.0080 	-.0::8
(1.83) 	(-0.63) 	(-0.60) 	(~.92)
FORCE 	-.0145 	-.0145 	-.0146 	-.0029
(-2.24) 	(-2.50) 	(-1.98) 	(~.18)
CANCEL 	.0206 	.0243 	.C2~ 3 	.01l9
(3.31 ) 	(4.33) 	(3.6 J) 	(0.76)
LD 	-.0237 	-.0:44 	-.0103 	.0224
(-3.80) 	(-4.35) 	(-1.45) 	( 1..39)
Chi
Square 	115.85 	191.24 	77.99 	46.73
29

.'. 	.
Appendix C - Ia ble 2
The Impact of the RMA La 90'S on Dealer Margins for Chevrolet Cars in 1978
Body Type
Regular
Malibu
Camara
Nova
Monte Carlo
MODza
Chevette
Sportvan
Corvette
Average 	Percentage 	Markup 	(Dealer
M:rgjn) 	Change 	For 	Areas 	with 	Positive
population Growth
3.69 0
3.41 0
•
8.17 0
•
2.23
4.07 0
7.67 00
5.15··
3.42
16.tl9°·
• significantly above zero at the 95% level.
~o significan tly above zero at the 99% level.
30