FTC Staff Comment Before the National Highway Traffic Safety Administration In the Matter of: Passenger Automobile Average Fuel Economy Standards Model Year 1978-88

BEFORE 	THE
NATIONAL 	HIGHWAY 	TRAFFIC 	SAFETY
Washington, 	D.C.
COiVlMISSIOI
_. 	APPROVED
ADMINISTRATION 	~;
-- 	. W6Y~378f£~ 	S]
In 	the 	Matter 	of: 	)
)
Passenger 	Automobile 	Average 	)
Fuel 	Economy 	Standards 	)
Model 	Year 	1987-88 	)
--------------)
Docket 	No. 	FE-85-01
COMMENTS 	OF THE 	BUREAUS 	OF 	COMPETITION,
ECONOMICS 	AND 	CONSUMER 	PROTECTION OF 	THE
FEDERAL 	TRADE 	COMMISSION
Washington, 	D.C. 	20580
Ma r ch 	2 6, 	19 86
On 	January 	16, 	1986, 	the 	National 	Highway 	Traffic 	Safety
Administration 	(NHTSA) 	released 	a 	Notice 	of 	proposed 	Rulemaking
("NPRM") 	in 	which 	it 	proposed 	that 	the 	Corporate 	Average 	Fuel
Economy 	(CAFE) 	standards 	for 	the 	1987 	and 	1988 	model 	years 	(MY)
be 	reduced 	from 	the 	planned 	27.5 	miles 	per 	gallon 	(mpg) 	to 	as 	low
as 	26 	mpg, 	the 	current 	CAFE 	standard. l 	In 	response 	to 	that 	NPRM,
the Bureaus 	of 	Consumer 	Protection, 	Competition, 	and 	Economics 	of
the 	Federal 	Trade 	Commission 	("FTC 	staff") 	submit 	the 	following
analysis 	of 	the 	short-term 	effects on 	consumer 	welfare 	and
automobile 	worker 	employment 	if 	the 	planned CAFE 	standard of 	27.5
mpg 	is 	retai ned. 2
1 	In 	the 	Matter 	of 	Passenger 	Automobile 	Average 	Fuel 	Economy
Standards 	Model 	Years, 	1987-88, 	Docket 	No. 	FE-85-01, 	51 	Fed. 	Reg.
2912-2923 	(1986) 	(hereinafter 	ci ted 	as 	"NPRM").
2 	These 	comments 	represent 	the views 	of 	the 	Bureaus 	of
Competition, 	Economics, 	and 	Consumer 	Protection and 	do 	not
necessarily 	represent 	the views 	of 	the 	Federal 	Trade 	Commission
or 	any 	individual 	Commissioner. 	The 	Commission has, 	however,
authorized 	the 	staff 	to 	submit 	these 	comments. 	Inquiries
regarding 	these 	comments 	should 	be 	directed 	to 	Richard 	Higgins,
Bureau 	of 	Economics, 	Federal 	Trade 	Commission, 	washington, 	DC~
20580.

This analysis 	is germane to question 17 in the NHTSA NPRM of
January 	16.3 	That 	questions 	reads: 	·What 	would 	be 	the 	likely
economic 	effects, 	i&...., 	effects 	on 	employment, 	car 	sales,
restrictions 	on 	consumer 	choice, 	etc., 	of 	alternative 	MY 1987-
1988 	passenger 	car 	fuel 	economy 	standards 	within 	the 	26 	mpg 	to
27.5 	mpg 	range?" 	As 	explained 	below, 	we 	conclude 	that 	raising
the 	actual 	CAFE 	standard 	to 	27.5 	mpg 	will 	have 	an 	adverse 	effect
on 	the 	welfare 	of 	American 	consumers 	and 	the 	0.5. 	economy.
Adoption 	of 	the 	more 	stringent 	standard 	would 	result 	in 	a
significant 	number 	of 	consumers 	purchasing 	low 	mileage 	cars 	at
prices 	higher 	than 	they 	would 	otherwise 	be 	and 	would 	also 	lower
employment 	in 	the 	automobile 	industry. 	These 	effects 	constitute
a 	cost 	against 	which 	the 	supposed 	benefits 	of 	the 	program 	should
be 	balanced. 	Because 	the 	benefits 	from 	the 	27.5 	mpg 	standard
appear 	small 	while 	the 	costs 	appear 	substantial, 	the 	FTC 	staff
urges 	NHTSA 	to 	lower 	the 	proposed 	standard 	for 	the 1987-88 	model
years.
These 	comments 	are 	divided 	into 	four 	sections. 	Section 	I
reviews 	the 	possible 	benefits 	that 	might 	follow 	from 	adoption 	of
the 	27.5 	mpg 	standard 	and 	concludes 	that 	the 	benefits 	are
relatively 	small. 	Section 	II 	develops 	a 	theoretical 	framework
for 	estimating 	the 	corresponding 	costs of 	the higher 	mileage
3 	NPRM 	at 	2919. 	The 	focus 	of 	these 	comments 	is on 	the 	consumer
welfare 	consequences 	of 	the 	proposed 	change 	in 	the 	CAFE 	standard,
all 	other 	factors 	being 	equal. 	We 	do 	not 	address 	the 	equitable
or 	other 	non-economic 	implications, 	if 	any, 	of 	different
manufacturers' 	varying 	degree 	of 	preparedness 	to 	meet 	the 	planned
27.5 	mpg 	standard, 	or 	of 	investment 	decisions 	that 	may 	have 	been
made 	in 	view 	of 	the 	planned 	standard. 	While 	such 	considerations
mayor 	may 	not 	be 	relevant 	to 	NHTSA's 	proposed 	rulemaking, 	the
affected 	manufacturers 	are 	in 	the 	best 	position to 	bring 	to 	light
the 	relevant 	facts 	and 	analysis.
2

standard, 	including 	its 	adverse 	effects 	on 	price, 	output, 	and
employment. 	Section 	III 	applies 	this 	theoretical 	model 	to 	the
facts 	of 	the 	auto 	industry 	and 	develops 	concrete 	estimates 	of 	the
total 	costs 	that 	will 	be 	borne 	by 	the 	public 	if 	the 	27.5 	mpg
standard 	is 	adopted. 	Finally, 	Section 	IV 	presents 	the 	conclusion
of 	the FTC staff, 	that the costs of 	moving to a higher mileage
standard 	will 	greatly 	exceed 	the 	benefits.
1. 	The 	Benefits 	of 	a 	27.5 	mpg 	CAFE 	Standa rd
When 	it 	was promulgated 	by 	Congress 	in 	1975,4 	the 	CAFE 	program
had 	three 	original 	rationales, 	each 	of 	which 	might 	arguably 	be
further 	served 	by 	increasing 	the 	mileage 	standard 	to 	27.5 	mpg.
These 	were 	(1) 	to 	complement 	price 	controls,S 	(2) 	to 	guard
against 	the 	risk 	that 	consumers 	would 	not 	respond 	significantly
to market forces and conserve fuel 	even in the absence of price
controls,6 	and 	(3) 	to 	improve 	national 	security.7
Subsequent 	developments 	call 	into question 	the 	relevance 	of
these 	concerns. 	First, 	CAFE 	regulation 	is 	no 	longer 	necessary 	as
a 	substitute for 	market 	forces 	because 	gas 	prices 	have 	been
4 	15 	u. S. C. 	§ § 1901, 	2001- 2012 •
5 	See, 	for 	example, 	the 	discussion 	of 	M. 	Weidenbaum, 	"U.S.
Should 	Scrap Law 	on 	Fuel 	Economy," 	Christian 	Science 	Monitor 	(May
22, 	1985) 	cited 	in 	Congressional 	Record 	(Senate) 	(July 	29, 	1985)
at 	S10239.
6 	See, 	for 	example, 	H. 	Rowan, 	·Cop-Out 	on 	Fuel 	Efficiency,"
Washington 	Post 	(July 	25, 	1985) 	cited 	in 	Congressional 	Record
(Senate) 	(July 	29, 	1985) 	at 	Sl0233.
7 	~, 	for 	example, 	remarks 	of 	Senator 	Durenberger 	Congressional
Record 	(Senate) 	(July 	29, 	1985) 	at 	Sl0235.
3

decontrolled 	and 	market 	forces 	reintroduced. 	Any 	protection 	from
higher 	gasoline 	prices 	afforded 	consumers 	by 	the 	domestic 	price
control 	program 	came 	to 	an 	end 	in 1981 	with 	its 	dismantling.
American 	consumers 	now 	purchase 	gasoline 	in 	a 	deregulated 	market.
Ironically, 	inflation-adjusted 	gasoline 	prices 	have 	actually
declined 	since 	the 	end 	of 	price 	controls. 8
Second, 	there 	is 	evidence 	that 	consumers 	in 	fact 	purchase
more 	small 	cars 	when 	gasoline 	prices 	increase. 	In Europe 	and
Japan, 	where 	gasoline 	prices 	are 	much 	higher 	than 	those 	in 	the
United 	states, 	consumers 	tend 	to 	purchase 	cars 	with 	much 	higher
gas 	mileage. 9 	In 	addition, 	U.~ 	consumers 	were 	in 	fact
purchasing more 	higher 	mileage 	cars 	in 	response 	to 	actual 	and
expected 	increases 	in 	gasoline 	prices 	even 	prior 	to 	adoption 	of
CAFE.lO 	Further, 	according 	to 	one 	recent 	study, 	the 	change 	in
average 	new 	car 	mileage 	between 	1970 	and 	1983 	was 	the 	same 	as 	it
would 	have 	been 	without 	the 	CAFE 	program. ll 	That 	is, 	as 	gasoline
prices 	rose 	gradually 	under 	the 	domestic 	price 	control 	program
and 	more 	rapidly 	following 	the 	end 	of 	price 	controls, 	consumers
responded 	to 	the 	high 	gasoline 	prices 	by 	purchasing 	more 	fuel-
efficient 	cars. 	The 	extent 	of 	these 	purchases 	was 	great 	enough
to 	render 	the 	CAFE 	mileage 	standards 	redundant.
8 	See 	D. 	Henderson, 	"The 	Economics 	of 	Fuel 	Economy 	standards,"
Regulation 	(January 	1985) 	at 	45-48 	(hereafter 	"Henderson").
9 	See 	R. 	Crandall, 	"Why 	Should We 	Regulate 	Fuel 	Economy 	at
All?", 	Brookings 	Review 	(Spring 	1985) 	at 	4 	(hereafter
"C ran da 11 ") •
10 ~.
11 	~ 	R. 	Crandall, 	T. 	Keeler, 	H. 	Gruenspecht, 	and L. 	Lave,
Regulating 	The 	Automobile, 	Brookings 	Institution 	(forthcoming).
4

Finally 	national 	security 	may 	also 	be 	a 	less 	pressing
concern 	now 	than 	when 	the 	CAFE 	rules 	were 	first 	introduced. 	The
considerations 	here 	are 	complex 	and 	not 	entirely 	consistent. 	On
the 	one 	hand, 	our 	level 	of 	imports 	is 	now 	less 	than 	before, 	and
so 	our 	national 	risk 	exposure 	is 	correspondingly 	less. 	RS. 	oil
consumption 	declined 	by 	16.7 	percent 	between 	1978 	and 	1984, 	for
example, 	and 	u.s. 	oil 	imports 	fell 	by 	almost 	42 	percent. 12 	On
the 	other 	hand, 	our 	risk 	exposure 	may 	still 	be 	undesirably 	high,
particularly 	inasmuch 	as 	a 	foreign 	oil 	cutoff 	may 	corne 	about 	for
political as 	well 	as 	economic 	reasons. 	In 	balancing 	these 	two
considerations, 	we 	believe 	that 	the 	use 	of 	market-determined
gasoline 	prices 	will 	be 	effective 	in 	reducing 	our 	exposure 	to
foreign 	pressure. 	We 	therefore 	believe 	that 	stricter 	CAFE
requirements 	will 	not 	be 	necessary.13
In 	sum, 	it 	does 	not 	appear 	that 	our 	national 	security
interests 	would 	be 	greatly 	furthered 	by 	adoption 	of 	that
standard. 	Further, 	in 	the 	current 	environment, 	consumers
presumably 	do 	not 	have 	to 	be 	protected from 	falling 	gasoline
prices 	by 	a 	more 	stringent 	CAFE 	standard. 	Even 	during 	the 	era 	of
rising 	gasoline 	prices, 	consumers 	responded 	by 	purchasing 	higher
mileage 	cars 	to an 	extent 	that 	may 	have 	rendered 	the 	CAFE 	program
redundant. 	In 	the 	absence 	of 	CAFE, 	the 	free 	market 	choices 	of
In 	making 	this 	judgement, 	we 	recognize 	that 	there 	is a
necessary 	delay 	between 	a 	rise 	in 	gas 	prices 	and 	a 	consumer
response 	in 	terms 	of 	fuel 	savings. 	That 	delay 	would 	cause 	some
interim 	harm 	in 	the 	event 	that 	there 	is 	another 	oil 	cutoff.
12 	~, 	R 	Shay, 	KComments 	of 	the 	Department 	of 	Commerce
[concerning 	CAFE]K 	(letter 	to 	the 	NHTSA 	Administrator, 	April 	30,
1985) 	at 	3.
13
5

consumers 	may 	well 	have 	resulted 	in 	a 	new 	car 	mileage 	virtually
identical 	to 	that 	mandated 	by 	CAFE. 	It 	appears, 	then, 	that
imposition 	of 	the 	27.5 	mpg 	standard 	would 	generate 	no 	substantial
benefits 	to 	either 	consumers 	or 	our 	national 	security 	interest.
II. 	A Model 	for 	Assessing 	The 	Costs 	of 	CAFE
Not 	only 	do 	the 	benefits 	from 	a 	more 	stringent 	CAFE 	standard
appear 	small, 	but 	the 	short-run 	costs 	to 	consumers 	could 	be
substantial. 	These 	costs 	would 	be 	generated 	by 	the mismatch 	in
demand 	and 	production 	caused 	by 	a 	CAFE-mandated 	reduction 	in 	the
production 	of 	low 	mileage 	cars 	and 	an 	increase 	in 	the 	production
of 	high-mileage 	cars. 	This 	mismatch 	between 	production and
demand 	would 	result 	in 	higher 	prices 	for 	low 	mileage 	cars
(because 	fewer 	of 	these 	cars 	will 	be 	produced 	than 	the 	market
would 	otherwise 	demand) 	and 	lower 	prices 	for 	high-mileage 	cars.
Some 	consumers who would 	have 	preferred low-mileage 	cars 	in 	the
absence 	of 	the 	CAFE 	standard mayor 	may 	not 	purchase 	the 	same 	car 	at
the 	higher 	price 	induced 	by 	the 	compliance 	with 	CAFE. 	In 	either
event, 	these 	consumers 	will 	suffer 	a 	loss 	in 	their 	well-being.
At 	the 	same 	time, 	the 	CAFE-induced 	diversion 	of 	resources 	to 	the
expansion 	of 	high-mileage 	car 	production would 	be 	resources 	whose
value 	to 	consumers 	would 	be 	greater 	in low-mileage 	car 	production
or 	elsewhere 	in 	the 	economy. 	Finally, 	the 	effect 	of 	the 	more
stringent 	CAFE 	standard may 	be 	to 	reduce 	total 	domestic
automobile 	production and 	result 	in 	unemployment 	of 	displaced
automobile 	workers. 	This 	loss 	to 	the 	economy would 	be measured
by 	the 	value 	of 	goods 	and 	services these 	displaced 	workers 	would
have 	produced 	absent 	unemployment.
6

To quantify 	these 	costs 	to 	consumers 	of 	a 	27.5 	mpg 	CAFE
standard, 	we have 	constructed 	a model 	to estimate 	the 	production
shifts, 	price 	changes, 	and 	employment 	effects 	that 	would 	result
from 	that 	standard. 	Because 	a 	complete 	discussion 	of 	our
estimation 	procedure 	is 	provided 	in Appendix 	I, 	here 	we 	simply
describe 	in 	general 	terms 	the 	nature 	of 	our 	approach. 	In 	the
following 	section 	we 	will 	apply 	this 	model 	in 	order 	to 	produce
concrete 	predictions 	about 	the 	effects 	of 	different 	regulatory
changes.
We 	consider 	a 	market 	in 	which 	high 	mileage 	(~, 	small)
cars 	and 	low 	mileage 	(l.&....., 	large) 	cars 	are 	produced. 	Small 	cars
are 	assumed 	to 	be 	those 	with 	an Environmental 	Protection 	Agency
("EPA") 	aggregate 	mileage 	(a 	weighted 	average 	of 	city 	and 	highway
mileage) 	that 	meet 	or 	exceed 	27 	mpg 	while 	large 	cars 	are 	those
below 	27 	mpg.14
Automobile 	suppliers 	are 	also 	divided 	into two 	groups:
those 	firms 	that 	are 	expected 	to 	meet 	or 	exceed 	the 	proposed 	CAFE
standard of 27.5 	mpg and those firms that are unlikely to meet
the 	more 	stringent 	standard without 	altering 	their 	present
practices 	and 	incurring 	substantial 	costs. 	The 	firms 	most 	likely
to 	fall 	into 	the 	latter 	category 	are GM 	and 	Ford, 	producers 	of
14 	While 	this 	mileage 	criteria 	is 	arbitrary, 	tentative 	analysis
by 	the 	FTC 	staff 	indicates 	that 	modest 	changes 	in 	the 	criteria
would 	have 	little 	effect 	on 	our 	estimates 	infra. 	We 	also 	note
that 	some 	"small 	cars" 	in 	our 	analysis 	might 	be 	popularly
considered 	as 	large 	cars 	and 	vice 	versa. 	This 	possibility 	may
have 	implications 	for 	our 	results. 	~ 	the 	discussion 	infra 	at
n. 	32. 	We 	adopt 	the 	mileage-based 	small-car/large-car
terminology 	for 	exposition 	purposes. 	/
7

both 	large 	and 	small 	cars. 	At 	the 	end 	of 	1985, 	GM 	had 	attained 	a
CAFE 	mileage 	rating 	of 	only 	25.8 	while 	the 	corresponding 	figure
for 	Ford 	was 	26.3 15
The 	remaining domestic firms--Chrysler, 	AMC, 	Volkswagen,
Honda, 	and 	Nissan--should 	have 	little 	difficulty 	in 	satisfying 	a
more 	stringent 	CAFE 	standard. 	These 	firms 	are 	primarily 	small
car 	producers. 	At 	the 	end 	of 	1985, 	they 	had 	all 	attained 	a 	CAFE
mileage 	rating 	that 	met 	or 	exceeded 	the 	proposed 	27.5 	mpg
standard.16 	The 	great 	majority 	of 	imports 	is 	also 	expected to
fall 	into 	this 	second 	group. 	In 	1983 	more 	than 	82 	percent 	of 	all
imported 	cars 	consisted 	of 	models 	attaining a 	mileage 	rating 	in
exce ss 	of 	27.5 	mpg.17
Against 	that 	background 	we 	construct 	a 	model 	of 	the 	demand
and 	supply 	conditions 	for 	the 	u.s. 	automobile 	market. 18 	As 	a
convenient 	benchmark we assume 	that 	the 	consumer 	demands 	for
large 	and 	small 	cars 	are 	distinct 	so 	that 	the demand 	for 	small
cars is independent of the price of large cars and the demand for
large 	cars 	is 	independent 	of 	the 	price 	of 	small 	cars. 	This
assumption 	is 	not 	inconsistent with 	automobile 	demand 	studies
which 	typically 	fail 	to 	detect 	any 	significant 	relationship
15 	These 	figures 	were 	obtained 	from 	GM 	and 	Ford.
16 	For 	Chrysler, 	its 	1985 	CAFE 	rating 	was 	27.6; 	for 	AMC, 	32.8;
for 	Volkswagen 	30.2; 	for 	Honda, 	33.9; 	and 	for 	Nissan 	29.4 	These
ratings 	were 	obtained 	directly 	from 	the 	various 	suppliers.
17 	See 	P. 	Hu 	and G. 	Roberts, 	RMotor 	Vehicle 	MPG 	and 	Market 	Share
Report, 	Model 	Year 	1983,R 	Oak 	Ridge 	National 	Laboratory 	(1984).
18 	A complete 	discussion 	of 	these 	assumptions 	and the 	relevant
sources 	in 	the 	economics 	literature 	can 	be 	found 	in 	Appendix 	I.
8

between the demand for 	large cars and the price of small cars {or
the 	demand 	for 	small 	cars 	and 	the 	price 	for 	large 	cars).19
Based 	on 	past 	studies 	of 	the 	demand 	for 	automobiles 	we
assume 	that 	a 	10 	percent 	increase 	in 	the 	small-car 	price 	will
generate a 20 percent decline in the quantity of small 	cars
demanded and that a 10 percent increase in the price of large
cars 	will 	lower 	the quantity 	of 	large 	cars 	demanded 	by 	30
percent. 20 	Most 	studies 	have 	found 	that 	consumer 	price 	sensitivity
to 	large 	and 	small-car 	prices 	is 	no 	greater 	than 	that 	assumed
here. 21 	The 	effect 	of 	our 	price 	sensitivity 	assumptions 	on 	the
estimated 	consumer 	cost 	of 	the more 	stringent 	CAFE 	standard 	is 	to
understate 	those 	costs 	if 	in 	fact 	price 	sensitivity 	is 	less 	than
assumed.
We assume that GM and Ford, 	the two producers that would be
affected 	by 	the 	27.5 	mpg 	standard, 	can, 	in 	the 	short 	run,
increase 	(up 	to 	a 	capacity 	constraint) 	or 	decrease 	the 	production
of 	small 	and 	large 	cars 	at 	constant 	variable 	unit 	costs. 	In
19 	~ 	the 	discussion 	in 	Appendix 	I. 	We 	discuss 	the 	sensitivity
of 	our 	results 	to 	this 	assumption 	infra. 	~ 	also 	n. 	14.
20 	These 	sensitivity 	estimates 	are 	those 	of 	J. 	Langenfeld 	and
M. 	Munger, 	"The 	Impact 	of 	Federal 	Automobile 	Regulations 	on 	Auto
Demand," 	unpublished 	draft, 	Federal 	Trade 	Commission 	(June 	1985).
21 	For 	example, 	one 	recent 	study 	(among 	others) 	provides 	lower
estimates 	of 	consumer 	price 	sensi tivi ty 	for 	a 	variety 	of 	car
types. 	The 	lowest 	estimate 	is 	for 	subcompacts, 	suggesting 	that 	a
10 	percent 	increase 	in 	price 	would 	reduce 	the 	number 	of
subcompacts 	demanded 	by 	only 	8 percent. 	The largest 	estimate 	is
for 	-luxury- 	cars, 	indicating 	that 	a 	ten 	percent 	price 	increase
would 	result 	in 	24 	percent 	fewer 	luxury 	cars 	demanded. 	See 	R.
Carlson, 	-Seemingly 	Unrelated Regression and 	the 	Demand 	for
Automobiles 	of 	Different 	Sizes, 	1965-75,- 	Journal 	of 	Business
(1978) 	at 	254-259.
9

particular, 	we assume 	that 	small-car 	production 	cannot 	be
increased 	during 	the 	next 	two 	years 	by 	more 	than 	30 	percent 	over
1985 	production 	levels. 	In 	the 	short 	run, 	GM and 	Ford 	can
increase 	the 	production 	of 	small 	cars 	by 	employing more 	worker
shifts and 	by 	reconfiguring 	some 	large-car 	plants 	for 	small-car
production. 	However, 	any 	dramatic 	increase 	in 	output 	in 	the
short-term 	is 	limited 	by GM's 	and 	Ford's 	ability 	to 	so
reconfigure 	large-car 	production 	facilities 	and 	by 	their 	ability
to 	acquire 	enough 	of 	the 	needed 	inputs, 	such 	as 	engines 	and
transmissions, 	to 	produce 	more 	small 	cars. 	In 	light 	of 	those
limitations, 	a maximum 	increase 	of 	30 	percent 	in 	small-car
production 	over 	a 	two-year 	period 	seems 	reasonable. 22 	Over 	a
longer 	time 	period, 	GM and 	Ford would 	build additional 	small-car
plant 	capacity 	and 	their 	input 	suppliers 	would also 	expand
production capacity to satisfy the input needs of GM and Ford at
an 	even 	larger 	small-car 	production 	rate. 	We 	also 	assume 	that
the 	two 	companies 	supply 	cars 	competitively 	given 	the 	CAFE
requi rements.
In response 	to the reduced production of large cars by GM
and 	Ford, 	the 	price 	of 	large 	cars 	will 	tend 	to 	rise. 	But 	as 	the
production 	of 	small 	cars 	rises, 	the 	small-car 	price 	will 	be
22 	The 	sensitivity 	of 	our 	results 	to 	the 	30 	percent 	constraint 	is
discussed 	infra 	at 	n. 	33.
10

reduced so that GM and Ford to sell 	the additional 	small 	cars
necessary 	to 	comply 	with 	CAFE.23
We expect 	that 	the shifts in production by GM and Ford and
the 	consequent 	change 	in 	prices 	will 	create 	incentives 	for 	the
small-car 	suppliers 	in 	the 	industry 	to 	alter 	their 	own
production. 	when 	the 	price 	of 	small 	cars 	declines, 	these
suppliers 	of 	small 	cars 	will 	tend 	to 	reduce 	their 	small-car
output. 	Thus, 	when 	the 	price 	of 	small 	cars 	declines 	as 	a 	result
of 	GM's 	and 	Ford's 	CAFE-induced 	production 	shifts, 	the 	reduction
in 	the 	small-car 	production 	of 	the 	remaining 	small-car 	suppliers
will 	ameliorate 	this 	price 	decline. 	To 	calculate 	the 	estimates
presented 	below, 	we assume 	that 	a 	one 	percent 	decline 	in 	the
23 	Alternatively, 	GM 	and 	Ford 	could 	opt 	to 	pay 	a 	penalty 	for 	any
shortfall 	between 	the 	attained 	CAFE 	rating 	and 	the 	27.5 	mpg 	CAFE
standard. 	Under 	its 	statutory 	authority, 	NETSA 	is 	required 	to
levy 	financial 	penalties 	in 	the 	event 	of 	a 	shortfall. 	See 	15
U.S.C. § 2008. 	To 	the 	extent 	that GM 	or 	Ford 	(or 	both) 	do 	rely 	on
the 	penalty 	option, 	the 	costs 	of 	CAFE 	presented 	here 	may 	be
overstated. 	However, 	the 	payment 	of 	the 	fine 	would 	still 	impose
losses 	on 	society 	similar 	to 	those 	in 	the 	text. 	Because 	the 	fine
is 	attributable 	to 	the 	sale 	of 	cars 	that 	do 	not 	meet 	the 	27.5 	mpg
standard, 	the fine would be similar to a tax on what we have
called 	large 	cars. 	Consequently, 	the 	price 	of 	large 	cars 	would
still 	rise 	and 	large-car 	production 	would 	fall. 	But 	because 	any
additional 	small-car 	sales 	would 	reduce 	the 	fine, 	small-car
output 	would 	rise 	and 	the 	small-car 	price 	would 	fall. 	Thus, 	a
fine 	payment 	still 	induces artificial 	production 	shifts 	similar
to 	those 	described 	in 	the 	text, 	although 	the 	magnitude 	of 	the
shifts 	may 	be 	different. 	These 	artificial 	production 	shifts
induced 	by 	the 	payment 	of 	the 	fine 	would 	result 	in 	social 	losses
identical 	in 	kind 	to 	those 	described 	in 	the 	text.
11

small-car 	price 	will 	result 	in 	a 	ten 	percent 	decline 	in 	the
output 	of 	small-car 	suppliers.24
Our 	analysis 	treats 	Chrysler 	and 	the 	imports of 	Japanese
cars 	somewhat 	differently 	from 	these 	other 	small-car 	suppliers.
Chrysler 	is 	now 	virtually 	in 	compliance 	with 	the 	proposed 	27.5
mpg 	standard. 	Nonetheless, 	Chrysler 	does 	produce 	some 	low
mileage 	(or 	what 	we 	have 	called 	large) 	cars. 25 	If 	the 	price 	of
24 	See 	D. 	Tarr 	and 	M. 	Morkre, 	"Aggregate 	Costs 	to 	the 	United
States 	of 	Tariffs 	and Quotas 	on 	Imports," 	(Federal 	Trade
Commission, 	December 	1984) 	at 	63, 	n. 	12. 	These 	authors 	infer 	the
output 	response 	of 	automobile 	producers 	to 	price 	changes 	from
previous 	studies 	assessing 	the 	relationship 	between 	U.S.
automobile 	prices and 	the 	share 	of 	domestic 	car 	purchases
accounted 	for 	by 	foreign 	imports which 	largely 	consist 	of 	small
cars. 	While 	we 	could 	have 	assumed, 	analogous 	to GM 	and 	Ford, 	that
small-car 	suppliers 	can 	reduce 	output 	at 	constant 	unit 	variable
costs, 	this 	cost 	behavior 	would 	lead 	to 	a 	complete 	short-run
withdrawal 	of 	these 	suppliers 	from 	the 	U.S 	market 	if 	the 	price 	of
small 	cars were 	to fall. 	In 	the framework 	developed here, 	GM 	and
Ford 	are 	willing 	to 	incur 	losses on 	small-car 	production to
comply 	with 	the 	CAFE 	standard 	in 	order 	to 	produce 	the 	now 	higher
priced 	(and 	thus 	more 	profitable) 	large 	cars. 	Consequently, 	GM
and Ford continue to produce small cars in order to 	sell large
cars. 	Because 	small 	car 	suppliers 	are 	not 	producing 	large 	cars,
large-car 	production 	is 	not 	a 	short-run 	option 	for 	these
suppliers. 	Confronted 	with 	a 	lower 	small-car 	price 	and 	constant
unit variable 	costs, 	these 	suppliers 	would 	minimize 	their 	short-
run 	losses 	by 	market 	place 	withdrawal. 	Because 	such 	a 	withdrawal
appears 	implausible, 	we 	instead assume 	in 	the 	text 	an 	output
response 	by 	small-car 	suppliers 	to a 	small-car 	price 	reduction
that, 	while 	large, 	still 	permits 	them 	to maintain a 	presence 	in
the 	o.s. 	market.
25 	While 	our 	definition 	of 	"large" 	cars 	may 	include 	some 	of
Chrysler's 	models 	conventionally 	thought 	of 	as 	"small" 	cars, 	our
large-car/small-car 	terminology 	is 	used only 	for 	expositional
simplicity. 	Nonetheless, 	the 	mileage-based 	definition 	of 	small
and large cars could affect our estimates of the costs of the
27.5 	mpg 	standard. 	~ 	the 	discussion 	infra 	at 	n. 	32.
12

small 	ca rs 	decl ines, 	any 	induced 	decl ine 	in 	Chrysl err s 	small-ca r
output 	would 	result 	in 	non-compliance 	with 	the 	27.5 	mpg 	standard.
We 	instead 	assume 	that 	Chrysler 	would 	maintain 	its 	large-car 	and
small-car 	production, 	remaining 	in 	compliance with 	the 	27.5 	mpg
standard.
With 	respect 	to 	U.S. 	imports 	of 	Japanese 	cars, 	we 	make 	two
alternative 	assumptions. 	In what 	we 	will 	refer 	to as 	simulation
(1), 	we 	assume 	that 	in 	the 	absence 	of 	the 	Voluntary 	Restraint
Agreement 	("VRA") 	Japanese 	imports 	would 	be 	considerably 	greater
than 	they 	are 	now. 	Under 	this assumption any 	small-car 	price
decline 	induced 	by 	CAFE 	compliance 	would 	have 	no 	effect 	on 	the
imports 	of 	small 	cars from 	Japan. 	Thus 	we 	are 	assuming 	that 	the
pent-up demand for Japanese cars is so strong that even at a
lower 	small-car 	price, 	the 	Japanese 	will 	continue 	to 	export 	to
the 	u.s. 	the 	maximum 	number 	of 	cars 	permitted 	by 	the 	VRA. 	In
simulation 	(2), 	we 	assume 	that 	the 	VRA 	is 	not 	an 	effective
constraint and 	that 	the 	CAFE-induced 	small-car 	price 	decline
would 	result 	in 	the 	same 	percentage 	reduction 	in Japanese 	imports
as 	for 	other 	non-Chrysler 	small-car 	suppliers. 	The 	two 	simula-
tions 	provide 	estimates 	that 	would 	bracket 	the 	likely 	social 	loss
associated 	wi th 	CAFE.
In 	summary, 	then, 	attainment 	of 	the more 	stringent 	CAFE
standard 	by GM and 	Ford 	would 	result 	in 	total 	large-car 	output
falling 	and 	total 	small-car 	output 	increasing. 	The 	net 	effect 	of
these 	production 	shifts 	could 	be 	to 	generate 	employment 	losses 	in
the 	U.S. 	automobile 	industry. 	The 	decline 	in 	large-car 	outp~t
would 	be 	accompanied 	by 	a 	higher 	large-car 	price 	while 	to 	a
13

lesser extent the price of small cars would tend to decline as
total 	small-car 	production 	expands.
As 	a 	result 	of 	these 	CAFE-induced 	changes 	in 	output 	and
prices, 	large-car 	consumers 	would 	experience 	losses 	because 	of
the 	higher 	large-car 	price. 	While 	small-car 	consumers 	would
experience 	some 	gains 	because 	of 	the 	somewhat lower 	small-car
price, 	on 	balance 	consumers 	would 	tend 	to 	experience 	losses 	from
imposition 	of 	the 	more 	stringent 	CAFE 	standard. 	Some 	of 	these
consumer 	losses 	would 	be 	offset 	by 	the 	higher 	profits 	earned 	by
U.S. 	producers 	on 	the 	sale 	of 	large 	cars. 	The 	remaining 	losses
are 	pure 	social 	deadweight 	losses. 	These 	could 	be 	avoided 	by
retaining 	the 	present 	26 	mpg 	CAFE 	standard.
III. 	Applying 	the 	Theoretical 	Model
We will 	new 	try 	to quantify the extent of those losses.
Based on 	the framework 	described above, 	Table 	I 	presents 	the
estimates 	of 	the 	price 	and 	output 	effects that 	would 	be 	generated
by 	a 	27.5 	mpg 	standard. 	In 	column 	I 	we 	report 	the 	actual 	1985
output and average prices of large and small cars sold in the
United 	States. 	Column 	2 	(simulation 	(1» 	contains 	the 	output 	and
price 	estimates 	that 	would 	result 	from 	the 	imposition of 	the 27.5
mpg 	standard, 	assuming 	that 	U.S. 	imports 	of 	Japanese 	small 	cars
do 	not 	decrease. 	Column 	3 	(simulation 	(2» 	reports 	the 	output
and 	price 	estimates 	that 	would 	result 	from 	the 	imposition 	of 	the
27.5 	mpg 	standa rd, 	assum ing 	that 	U. s. 	impor ts 	of 	Japane se 	sm all
cars 	fall 	when 	the 	price 	of 	small 	cars 	falls.
The 	price 	of 	small 	cars 	declines 	in 	simulation 	(1) 	by 	4.3
percent 	and 	by 2.1 	percent 	in 	simulation 	(2). 	The 	extent 	of 	the
14

Table 	I
The 	Effect 	of 	Raising 	the 	CAFE 	Standard
from 	26 	mpg to 27.5 	rnpg
( 1)
1985
Act ua1
(2 )
Simulation 	(1)1
(3 )
Simulation 	(2)2
Average
Large 	Car
Price 	. 	$10,991 	$13,368 	$13,368	.
GM-Ford
Large 	Car
Sales 	3,593,711 	1,832,317 	1,832,317
Total
Large 	Car
Sales 	3,965,256 	2,203,861 	2,203,861
Average
Small 	Car
Price: 	$9,041 	$8,651 	$8,851
GM-Ford
Small 	Car
Sales 	3,073,819 	3,995,965 	3,995,965
Total
Small 	Car
Sales 	7,067,082 	7,717,373 	7,374,052
Total
Sales
(units) : 	11,032,338 	9,921,234 	9,577,913
Total
Domestic
Company
Sales
(units) : 	8,194,222 	7,221,819 	7,288,349
1 U.S. 	imports 	of 	Japanese 	srn all 	cars 	are 	unchanged 	from 	1985
level s.
2 U.S. 	imports 	of 	Japanese 	small 	cars 	decline 	when 	the 	small-car
price 	falls.
II
f
15

small-car 	price 	decline 	is 	less 	in 	simulation 	(2) 	because 	in 	this
simulation 	all 	non-Chrysler 	small-car 	suppliers 	(including
offshore 	Japanese 	producers) 	respond 	to 	the 	lower 	small-car
prices 	by 	reducing 	output 	of 	small 	cars. 	Thus, 	to 	sell 	any 	given
number 	of 	small 	cars 	to 	comply 	with 	the 	CAFE 	standard, 	GM 	and
Ford 	need 	not 	experience 	as 	sharp 	a 	decline 	in 	the 	small-car
price. 	In 	simulation 	(1), 	because 	u.s. 	small-car 	imports 	from
Japan do not fall as the small-car price falls, 	GM. and Ford must
further 	reduce 	the 	small-car 	price 	to 	induce 	the larger 	consumer
purchases 	of 	small 	cars 	necessary 	to 	comply 	with 	the 	27.5 	mpg
standard.
In 	both 	simulations, 	GM and 	Ford 	increase 	small-car 	output
until 	full 	capaci ty 	is 	reached 	(3,995,965 	small 	cars).
Nonetheless, 	total 	small-car 	output 	is 	4.7 	percent 	higher 	in
simulation 	(1) 	than 	in 	simulation 	(2). 	The 	difference 	here 	is
again 	explained 	by 	our 	assumption 	in 	each 	simulation 	regarding
the 	output 	response 	of 	offshore 	Japanese 	suppliers 	to 	the
reduction 	in 	the 	small-car 	price.
In 	both 	simulations, 	compliance 	with 	the 	more 	stringent 	CAFE
standard 	leads 	to 	a 	significant 	decrease 	in 	large-car 	sales. 	GM
and 	Ford 	reduce 	large-car 	sales 	by 	more 	than 1.7 	million 	units--
almost 	one-half 	of 	their 	current 	output 	of 	large 	cars. 	The
equality 	of 	the 	large-car 	reduction 	in 	both 	simulations 	is
explained 	by 	the 	small-car 	capacity 	constraint 	on GM. 	and 	Ford.
In 	each 	simulation, 	Ford and GM 	increase 	small-car 	production 	as
much 	as 	capacity 	permits. 	Thus, 	as 	noted 	above, 	they 	produce 	the
same 	number 	of 	small 	cars 	under 	either 	scenario. 	The 	necessity
of 	complying with 	the 	27.5 	mpg 	standard 	insures 	that 	the 	number
16

of large 	cars produced by GM and Ford also is the same in both
simulations. 26 	As 	a 	consequence 	of 	the 	decline 	in 	large-car
output, 	the 	price 	of 	large 	cars 	increases 	in 	both 	simulations 	by
21.6 	percent.
These 	simulations 	indicate 	that 	the 	short-run effect 	of
increasing 	the 	CAFE 	standard 	to 	27.5 	mpg 	would 	be 	a 	decrease 	in
the 	production 	of 	all 	cars 	in 	the 	United 	States. 	In 	simulation
(1), 	total 	sales 	drop 	by 	972,403 	units 	while 	the 	corresponding
decl ine 	in 	si mul ati on 	(2) 	is 	910,873 	uni ts. 27 	The 	decl ine 	in
both 	simulations 	occurs 	because 	capacity 	constraints 	on 	small-car
production 	by 	GM 	and 	Ford 	limit 	their 	short-run ability 	to
respond 	to 	the 	stricter 	CAFE 	standard 	by 	increasing 	their
production 	of 	small 	cars. 	Thus, 	the 	standard 	must 	be 	met 	by
substantial 	large-car 	output 	reductions.
The 	decline 	in 	u.s. 	production 	would 	lead 	to 	the 	substantial
displacement 	of 	workers 	in 	the 	automobile and 	auto 	parts 	indus-
26 	If, 	given the number of 	small cars produced by GM and Ford,
these 	two 	suppliers 	increased 	the 	number 	of 	large 	cars 	produced
beyond that 	in Table 	I, 	GM and Ford would not be in compliance
with 	CAFE. 	If 	they 	instead 	produced 	fewer 	large 	cars, 	GM 	and
Ford 	would 	not 	be 	profit-maximizing because 	they 	would not 	be
taking 	full 	advantage 	of 	the 	rise 	in 	large-car 	prices.
27 	The 	decline 	is 	greater 	in 	simulation 	(1) 	because 	the 	small-car
price 	decline 	(and 	thus 	the 	domestic 	supplier 	reduction 	in 	small-
car 	output) 	is 	greater 	than 	in 	simulation 	(2). 	Again, 	the
difference 	is 	explained 	by 	the 	differing 	assumptions 	regarding
the 	output 	response 	of 	offshore 	Japanese 	small-car 	suppliers to
small-car 	price 	declines.
17
jI
I
1!
I
I

tries. 	Employment 	in 	simulation 	(1) 	declines 	by 	137,147 	while
that 	in 	simulation 	(2) 	falls 	by 	130,809. 28
Clearly, 	this 	employment 	loss 	is 	likely to 	be 	temporary 	as
these 	displaced 	workers 	find 	new 	employment 	(or 	auto makers 	build
new 	sm all-car 	capaci ty). 	Howeve r, 	while 	these 	employees 	are
engaged 	in 	seeking 	new 	jobs, 	the 	loss 	to 	the 	u.s. 	economy 	from
this 	temporary 	unemployment 	can 	be 	estimated 	by 	considering 	the
likely 	duration 	of 	unemployment 	and 	the 	value 	of 	the 	output 	these
workers would 	have 	produced absent 	the 	CAFE-induced 	unemployment
spell. 	In 	simulation 	(1), 	the 	output 	loss 	to 	the 	economy 	is 	$2.1
billion 	while 	that 	for 	simulation 	(2) 	is 	$2.0 	billion. 29
28 	To 	calculate 	the 	unemployment 	changes 	we 	use 	data 	from 	the
Congressional 	Budget 	Office 	that 	provides 	the 	additional 	hours 	of
work 	required 	to 	produce 	an additional 	automobile 	in 	this
country. 	See 	Domestic 	Content 	Legislation 	and 	the 	u.s.
Automobile 	Industry, 	Subcommittee 	on Trade 	of 	the House 	Committee
on 	Ways 	and 	Means 	(August 	16, 	1982) 	at 	34-48. 	From 	these 	we
compute 	a 	coefficient 	that 	shows 	the 	change 	in the number 	of 	jobs
that result from a 	change of a 	given size in the number of
automobiles 	produced. 	We 	have 	computed such 	a 	coefficient 	for
both 	small 	and 	large 	cars. 	Multiplying 	these 	coefficients 	by 	the
change in the sales of domestic large and small cars, 	we arrive
at 	estimates for 	the 	employment 	changes 	brought 	about 	by 	the
adjustment 	to 	the 	new 	CAFE 	standards.
The 	employment 	effect 	is 	greater 	than 	might 	appear 	based 	on
the 	change 	in 	total 	automobile 	production 	alone 	because 	the
production 	of 	large 	cars 	involves 	more 	labor 	than 	the 	production
of 	small 	cars 	as 	suggested 	by 	the 	Congressional 	Budget 	Office
data.
29 	These 	output 	losses 	are 	based 	upon an 	unemployment 	spell 	in
the 	automobile 	industry 	of 	20 	weeks 	and 	an 	average annual 	wage
(including 	benefits) 	of 	$40,360 	per 	automobile 	production 	and
parts 	worker. 	These 	1985 	data 	were 	provided 	by 	the 	Department 	of
Labor. 	The 	unemployment 	spell 	used 	is 	not 	the 	average
experienced 	by 	all 	unemployed automobile 	workers. 	That 	average
would 	incl ude 	workers 	who are 	temporarily 	rather 	than permanently
unemployed 	in 	the 	automobile 	industry and 	who, 	therefore, 	may
experience 	a 	relatively 	brief 	period 	of 	unemployment. 	Because
(footnote 	continued)
18

These 	output 	losses 	are 	not 	the 	only 	losses 	society 	bears 	as
a 	consequence 	of 	the 	more 	stringent 	CAFE 	standard. 	Obviously,
consumers 	are 	made 	worse 	off 	when 	the 	price 	of 	cars 	rises, 	and
are made better off when the price of 	cars falls. 	One measure of
this 	change 	in 	consumer 	well-being 	is 	the 	change 	in 	consumer
surplus. 30 	The 	gains 	or 	losses 	in 	consumer 	surplus, 	however, 	will
be 	offset 	in 	part 	by 	the 	losses 	or 	gains 	to 	producers, 	measured
as 	profits. 	The 	sum 	of 	the 	changes 	in 	consumer 	surplus 	and 	the
change 	in 	producer 	profits 	(or 	producer 	surplus), 	when 	combined
with 	the 	output 	losses 	generated 	by 	the 	CAFE-induced 	unemploy-
ment, 	represents the total loss imposed on the public as a conse-
quence 	of 	the 	higher 	CAFE 	standard.
Our 	calculations 	of 	consumer 	and 	producer 	surplus 	changes
for 	each 	of 	the 	demand 	scenarios 	are 	presented 	in 	Table 	11.31
(footnote 	continued)
workers 	who 	are 	permanently 	displaced 	in the 	automobile 	industry
may 	experience 	a more 	extended 	unemployment 	spell, 	we 	sought 	to
minimize 	the 	effect 	of 	the 	temporarily 	unemployed on our 	duration
estimate. 	We 	used 	the 	unemployment 	spell 	experienced 	by 	those
workers 	at 	the 75th 	percentile 	of 	all 	unemployed 	workers 	when
arrayed 	by 	the 	duration 	of 	unemployment.
30 	A decline 	in 	consumer 	surplus 	is 	measured 	as 	the 	increase 	in
the price paid by consumers who purchase cars at the 	old and new
prices plus 	the 	difference 	between what 	consumers 	who only
purchase cars at the lower price would be willing to pay and that
lower 	price. 	Similarly, 	an 	increase 	in 	consumer 	surplus 	is
measured as the decrease in the price paid by car consumers at
both 	the 	old 	and 	new 	prices 	plus 	the 	difference 	between what
consumers who only 	purchase 	cars at 	the 	lower 	price 	would 	be
willing 	to 	pay 	and 	that 	lower 	price. 	Thus, 	a 	rise 	in 	car 	prices
will 	always 	result 	in 	a 	consumer 	surplus 	decline, 	and 	vice 	versa.
31 	See 	Appendix 	II 	for 	the 	details 	of 	this 	calculatjon.
19
I
I

Table 	II
The 	Total 	Cost 	to 	Society 	of 	Raising 	the 	CAFE
Standard 	from 	26 	mpg 	to 	27.5 	mpg
(billions 	of 	dollars)
A. 	Output 	Change
Due 	to
Increased
Unemployment
B. 	Deadweight
Social 	Loss 	(-)
or 	Gain 	(+)
Simulation 	(1)1
-$2.129
-$0.925
Simulation 	(2) 2
-$2.031
-$1.422
Consumer 	Surplus
Change 	-$4.186 	-$5.688
Producer 	Surplus
Change 	+$3.261 	+$4.266
Total 	Loss 	(-) 	or
Gain 	(+ ) 	(A+B) 	-$3.054 	-$3.453
I 	u.S. 	imports 	of 	Japanese 	small 	cars 	are 	unchanged 	from 	1985
levels.
2 u. S. 	imports 	of 	Japanese 	small 	cars 	fall 	when 	the 	sm all-ca r
price 	declines.
20

In 	both 	simulations, 	the 	rise 	in 	the 	price 	of 	large 	cars
will 	tend 	to 	reduce 	consumer 	surplus 	while 	the 	fall 	in 	the 	price
of 	small 	cars 	will 	tend 	to 	increase 	consumer 	surplus. 	The 	net
effect 	is 	a 	decline 	in 	consumer 	surplus 	of 	$4.2 	billion 	in
simulation 	(1) 	and 	$5.7 	billion 	in 	simulation 	(2). 	Because 	the
large-car 	price 	increase 	is 	identical 	in 	both 	scenarios, 	the
decline 	in 	consumer 	surplus 	is 	the 	same 	for 	each 	set 	of 	large-car
consumers. 	However, 	because 	the 	small-car 	price 	decline 	is
greater 	in 	simulation 	(1), 	small-car 	consumers 	in 	this 	simulation
experience 	a 	larger 	gain 	in 	consumer 	surplus 	than 	those 	in
simulation 	(2). 	As 	a 	consequence, 	the 	net 	consumer 	surplus 	loss
in 	simulation 	(1) 	is 	less 	than 	that 	in 	simulation 	(2).
United 	States 	producers 	in 	both 	scenarios 	experience 	a 	rise
in 	profits 	or 	producer 	surplus 	on 	large-car 	sales 	because 	the
large-car 	price 	rises. 	However, 	they 	experience 	a 	decline 	on 	the
surplus 	they 	receive 	on 	small-car 	sales 	because 	the 	price 	of
these 	cars 	falls. 	Because 	the 	rise 	in 	large-car 	price 	is 	the
same 	in 	both 	simulations 	while 	the 	small-car 	price 	decline 	is
greater 	in 	simulation 	(1), 	the 	U.s. 	suppliers 	in 	simulation 	(1)
gain 	$3.3 	bill ion 	whil e 	in 	sim ula ti on 	(2) 	they 	gain 	$4.3 	bill ion.
In 	simulation 	(1), 	the 	difference 	between 	the 	consumer 	loss 	and
the 	domestic 	producer 	gain 	is 	$0.9 	billion 	while 	that 	in
simulation 	(2) 	is 	$1.4 	billion. 	In 	each 	simulation, 	this
difference 	is 	the 	"deadweight" 	loss 	to 	society 	from 	imposition 	of
the 	more 	stringent 	CAFE 	standard. 	The 	deadweight loss--which 	is
the 	consumer 	loss 	not 	captured 	by 	domestic 	producers 	in 	the 	form
21

of 	higher 	profits32--represents 	a 	cost 	to 	society 	that 	could 	be
avoided 	by 	lowering 	the 	proposed 	CAFE 	standard. 	When 	the
deadweight 	loss 	is 	combined 	with 	the 	unemployment-related losses,
the 	total 	cost 	of 	~he 	27.5 	mpg 	standard 	to 	society 	is 	$3.0
billion 	in 	simulation 	(1) 	and 	$3.5 	billion 	in 	simulation 	(2).33
32 	These 	calculations 	exclude 	the 	changes 	in 	profits 	earned 	by
non-U.S. 	producers 	because 	those 	changes 	will 	be 	borne 	largely 	by
non-U.S. 	citizens. 	Because 	any 	profits 	earned 	by 	foreign
suppliers 	are 	not 	included 	in 	the 	deadweight 	loss 	calculation,
this 	loss 	would 	be 	higher 	if 	contrary 	to what we 	assumed 	the
demand for 	small cars depends on the price of large cars and
vice 	versa. 	As 	a 	result 	of 	the 	CAFE-induced 	increase 	in 	the
large-car 	price, 	some would-be 	large-car 	consumers 	will 	opt
instead 	for 	small 	cars, 	assuming 	this 	demand 	interdependency.
Consequently, 	the 	price 	of 	small 	cars 	would 	rise 	and 	the 	profits
of 	U.S. 	and 	foreign 	small-car 	suppliers 	would 	increase. 	Because
of 	the 	high 	share 	of 	small-car 	sales 	accounted 	for 	by 	foreign
suppliers 	and 	because 	the 	increase 	in 	foreign 	supplier 	profits 	is
not included as an offset to the consumer losses experienced as a
result 	of 	the 	higher 	prices, 	the 	deadweight 	losses 	if 	there 	is 	a
demand 	interdependence 	would 	be 	greater 	than 	those 	in 	Table 	II.
Because 	we 	have 	defined 	small 	and 	large 	cars 	solely 	in 	terms 	of
gas 	mileage, 	low 	mileage 	(high mileage) 	cars may 	in 	fact 	be
viewed 	by 	consumers 	as 	small 	(large) 	cars. 	Thus, 	our 	definition
of 	large 	cars 	may 	in 	fact 	include 	cars 	that 	consumers 	consider 	as
small. 	When 	the 	price 	of 	our 	mileage-defined large 	cars
increases, 	consumers 	(who 	consider 	those 	low 	mileage 	cars 	as
small) 	may 	opt 	to 	purchase 	our 	mileage-defined 	small 	cars 	(many
of 	which 	consumers 	may 	also 	regard 	as 	small). 	Thus, 	if 	our
mileage-based 	definition 	of 	small 	and 	large 	cars 	does 	not
correspond 	to 	what 	consumers 	consider 	small 	and 	large, 	our 	car
demands 	would 	be 	interdependent 	and our 	deadweight 	loss 	estimates
would 	be 	understated. 	Because 	our 	consumer 	price 	sensitivity
estimates 	appear 	to 	be 	based 	on more 	conventional 	definitions of
large 	and 	small 	cars 	(see 	note 	21), 	we 	believe 	these
possibilities 	are 	likely.
33 	The 	total 	losses 	to 	U.S. 	producers 	and 	consumers 	would 	be
smaller 	if GM and 	Ford 	could 	in 	the 	short-term 	expand 	small-car
production 	by 	more 	than 	th~ assumed 	30 percent 	increase 	over 	1985
levels. 	For 	example, 	if 	we 	had 	assumed 	that 	the GM's 	and 	Ford's
small-car 	output 	could 	rise 	by 	as 	much 	as 	40 	percent 	over 	1985
levels, 	the 	total 	loss 	to 	u.S. 	producers 	and 	consumers 	in
simulation 	(1) 	would 	have 	been 	$1.8 	billion 	while 	the
corresponding 	loss 	in 	simulation 	(2) 	would 	have 	been 	$2.3
billion. 	While 	lower 	than 	the 	losses 	in 	Table 	II, 	the 	CAFE-
induced 	losses 	still 	remain 	substantial.
22

IV. 	Conclusion 	and 	Recommendations
The 	preceding 	analysis 	indicates 	that 	in 	the 	short-term, 	the
costs borne by society if the 27.5 mpg standard is imposed would
be 	substantial. 	The 	price 	of 	large 	cars 	may 	rise 	by 	as 	much 	as
22 	percent 	as 	large-car 	production 	drops 	by 	1.7 	million 	units.
While 	small-car 	production 	may 	rise 	by 	an 	amount 	between 	350,000
and 	670,000 	units, 	total 	domestic 	car 	production 	nonetheless 	may
fall 	by 	more 	than 900,000 	units. 	The 	short-term 	employment
effects 	are 	substantial: 	over 	130,000 	jobs 	in 	the 	domestic
automobile 	industry 	will 	disappear. 	OVerall, 	the 	sum 	of 	the
deadweight 	loss 	to 	consumers and 	producers 	and 	the 	output 	losses
caused 	by 	the 	temporary 	unemployment 	generated 	by 	the higher 	CAFE
standard would range between a low 	of 	$3.0 billion and a high of
$3.5 	billion. 34 	Because 	these 	losses 	appear 	substantial 	and
because 	the 	benefits 	from 	a more 	stringent 	CAFE 	standard 	appear
non-existent, 	the 	FTC 	staff 	urges 	the 	NHTSA 	to 	retain 	the 	current
standard 	of 	26 	mpg.
34 	In the longer 	term, 	GM and Ford may adapt to the more
stringent 	standard 	by 	altering 	the 	construction 	of 	their 	large
and 	small 	cars 	so 	as 	to 	attain 	the 	27.5 	mpg 	standard 	in 	a 	less
costly 	fashion. 	Consequently, 	the 	price 	of 	large 	cars 	may 	fall
somewhat 	and 	the 	price 	of 	small 	cars may 	increase somewhat from
their 	projected 	levels 	in 	Table 	I. 	As 	a 	result, 	the 	costs 	to
U.S. 	consumers 	and 	U.S 	producers 	of 	the 	CAFE 	standard 	in 	the
longer term may abate but may not disappear. 	The mix and
characteristics 	of 	both 	large 	and 	small 	cars 	that 	in 	the 	longer
term would be induced by the 27.5 mIlg standard may still differ
from 	those 	which 	consumers 	would have 	selected absent 	the more
stringent 	standard. 	If 	this 	were 	the 	case, 	the 	full 	loss 	to 	U.s.
consumers 	and 	producers 	would 	be 	the 	short-term 	losses 	in Table
II 	plus 	these 	continuing 	long-term 	losses. 	Alternatively, 	if 	the
longer-term 	mix 	and 	characteristics of 	large 	and small 	cars 	are
identical 	to 	those 	which 	consumers 	would 	have 	chosen 	absent 	the
27.5 	mpg 	standard, 	then 	the 	short-term 	losses 	in 	Table 	II 	are 	the
full 	losses 	from 	the 	higher 	CAFE mileage 	standard.

Appendix 	I
Estimating 	the 	Price 	and 	Output 	Effects 	of 	CAFEI
In 	this 	appendix, 	we 	describe 	the 	model 	used 	to 	determine
the 	price 	and 	output 	effects 	of 	raising 	the 	actual 	CAFE 	standard
to 	27.5 	mpg. 	We 	first 	describe 	the 	various 	supply 	and 	demand
relationships 	assumed, 	and 	then 	describe 	the 	sources 	for 	the
parameters 	used 	to 	make 	the 	necessary 	calculations.
A. 	The 	Derivation 	of 	the 	Model
As 	stated 	in 	the 	text 	of 	the 	comment, 	we 	begin 	with 	demand
curves 	for 	the 	two 	types 	of 	cars 	(large 	and 	small). 	Consistent
with 	recent 	research,2 	we 	assume 	a 	constant 	elasticity
specification 	for 	each 	demand:
(AI) °L = 	al 	PL fl 	Ps bl
( A2) 	OS 	a2 	Ps f2 PL b2
=
where °L = 	the 	total 	sales 	of 	large 	cars 	in 	the 	U. 	s. ,
PL = 	the 	average 	price 	of 	large 	cars 	in 	the 	U. 	s. ,
Os 	= 	the 	total 	sales 	of 	small 	cars 	in 	the 	U. 	s. ,
Ps 	= 	the 	average 	price 	of 	small 	cars 	in 	the 	U. 	s. ,
fl 	= the own price elasticity of demand for 	large cars,
eLL'
bl 	= the 	cross-elasticity 	of 	demand 	for 	large 	car
quantity 	with 	respect 	to 	small 	car 	price
e LS '
1 	The 	primary 	author 	of 	this 	and 	the 	follO\Ting 	Appendix 	is
Robert 	P. 	Rogers.
2 	See 	J. 	Langenfeld, 	"Federal 	Automobile 	Regulation," 	Ph.D.
dissertation 	(Washington 	University, 	St. 	Louis 	(May 	1983)) 	and
J. 	Langenfeld 	and 	M. 	Munger, 	"The 	Impact 	of 	Federal 	Automobile
Regulations 	on 	Auto 	Demand" 	(Unpublished 	draft, 	Federal 	Trade
Commission, 	(June 	1985)).
i
II
I
iI

f2 	= 	the 	own 	price 	elasticity 	of 	demand 	for 	small 	cars,
eSS'
b2 	= the 	cross-elasticity 	of 	demand 	for 	small 	car
quantity 	with 	respect 	to 	large 	car 	price
e SL ·
Our 	calculations 	can 	be 	best 	handled 	by 	using 	the 	inverses 	of
these 	functions:
(A3 ) 	PL = gO °L gl 	QS g2
(A4) 	Ps 	hO 	Os 	hI 	OL h2
=
((-e SS ) 	/ 	M) 	(eLS / 	M)
where 	gO 	= 	al 	a2
gl 	= 	(e SS 	) 	/ 	M
g2 	= 	(-eLS 	) 	/ 	M
( (-eLL 	/ 	M) 	(eSL 	/ 	M)
hO 	= 	a2 	al
hI 	= 	(eLL) 	/ 	M
h2 	= 	(-eSL) 	/ 	M
and 	M = (e SS 	* eLL 	- 	eLS 	* e SL 	) •
The 	two 	largest 	American 	companies, 	GM 	and 	Ford, 	are 	assumed
to 	maximize 	profits 	subject 	to 	the 	product 	mix 	determined 	by 	the
CAFE 	regulation 	and 	to 	a 	long-run 	zero 	economic 	profit 	constraint
imposed 	by 	the 	possibility 	of 	entry. 	We 	assume 	that 	in 	1985
such 	a 	zero 	profit 	equilibrium 	was 	reached. 	Further, 	we 	assume
GM 	and 	Ford 	are 	prevented from 	increasing 	small 	car 	output 	beyond
a 	capacity 	limit 	which 	is 	30 	per 	cent 	greater 	than 	1985 	output.
The position of 	these two firms, 	then, 	can be summarized by the
following 	equation 	system: 3
3 	For 	analytical 	ease, 	we treat GM and Ford as if they were a
single 	firm 	(GM-Ford). 	/
2

(AS) 	L(PL 	CL) 	+ 	S(PS 	- 	CS) 	= 	o.
(A6) 	L = R*S
(A7) 	S.s. SC,
where 	L = the GM-Ford sales of large cars
S = 	the 	GM-Ford 	sales 	of 	small 	cars,
CL = the 	GM-Ford 	average 	(and 	marginal) 	cost 	of 	producing
large 	cars 	(assumed 	equal 	to 	the 	1985 	price 	of 	large
ca r s) ,
the 	GM-Ford 	average 	(and 	marginal) 	cost 	of 	producing
small 	cars 	(assumed 	equal 	to 	the 	1985 	price 	of 	small
ca rs) ,
R = the 	CAFE 	determined 	large 	car-small 	car 	ratio, 	and
SC = the 	small 	car 	capacity 	limit 	(at 	1.3 	times
1985 	sales).
Equation 	(AS) 	is 	the 	profit 	function; 	equation 	(A6) 	is 	the 	CAFE
constraint; 	and 	inequality 	(A7) 	describes 	the 	capacity 	limit.
The 	prices 	in equation 	(AS), 	PL and 	PS, 	are 	determined 	by 	the
above-described 	inverse 	demand 	equations 	(A3) 	and 	(A4). 	The
ratio, 	R, 	in 	equation 	(A6) 	is 	a 	function 	of 	the 	CAFE 	regulation
and 	the 	mpg's 	of 	the 	car s 	in 	the 	GM-Ford 	fleet. 	The 	CAFE
regulation 	requires 	that 	the 	composite 	mileage 	(K) 	of 	all 	the 	cars
that 	a 	company 	sells 	be 	calculated 	in 	the 	following 	fashion:
K = (L + 	S) / « 	L / 	ML ) 	+ 	(S / 	MS 	)),
where 	ML and 	MS = the 	actual 	average 	miles 	per 	gallon 	respectively
of 	large 	and 	small 	cars.
The 	right-hand 	side 	of 	this 	equation 	is 	the 	weighted 	harmonic
mean of 	actual 	fleet mileage and must meet or exceed the 27.S 	mpg
3
II

standard. 	Following 	Kwoka,4 	this 	equation 	can 	be 	expressed 	in
terms 	of 	the 	CAFE 	required 	ratio 	between 	small 	and 	large 	car
sales 	represented 	by 	R in 	equation 	(A6) 	by 	letting 	K = 27.5.
(A6) 	L 	= 	R'*S 	where
(A6a) 	R = 	«M S - 	K) *ML) / «K 	- 	ML) *MS) .
We 	also assume 	that 	the 	cost 	functions 	for 	GM-Ford 	are 	of
the 	constant 	returns 	type. 	This 	means 	that 	regardless 	of
output, 	unit 	costs 	remain 	at 	the 	same 	level, 	Cs 	and 	CL.
Therefore 	the 	firms 	can 	increase 	output 	without 	changing 	their
average 	per 	uni t 	cost. S
With 	respect 	to 	the 	remaining 	suppliers 	(the 	fringe), 	we
represent 	their 	supply 	response 	functions 	for 	the 	outputs 	of
large 	and 	small 	cars 	as 	follows:
(A8) 	FL 	= FL(.),
(A9) 	FS 	= 	FS{.),
where 	FL 	and 	FS 	= the 	fringe 	sales 	of 	large 	and 	small
cars 	respectively.
The 	fringe 	consists 	of 	domestic 	car 	suppl iers 	(Chrysler, 	AMC
Volkswagen, 	Honda, 	and 	Nissan) 	and 	suppliers 	of 	imports 	to 	the
United 	states. 	with 	the 	exception 	of 	Chrysler, 	these 	suppliers
4 	See 	J. 	Kwoka, 	"The 	Limits 	of 	Market-Oriented 	Regulatory
Techniques: 	The 	Case 	of 	Automotive 	Fuel 	Economy,· 	Ouarterly
Journal 	of 	Econom ics « 	(Novembe r 	1983) 	at 	69S-7 04.
5 	There 	is 	some 	empirical 	work 	that 	supports 	making 	this
assumption. 	See 	A. 	Friedlaender, 	A.. 	Winston, 	and 	K. 	Wang,
"Costs, 	Technology, 	and 	Productivity 	in 	the 	U. 	S. 	Automobile
Industry," 	~ 	Journal 	of 	Economics, 	(Winter 	1982) 	at 	1-20 	.
4

are 	predominantly 	small 	car 	suppliers. 6 	While 	Chrysler 	does
produce 	a 	greater 	proportion 	of 	large 	cars 	than 	the 	other 	fringe
suppliers, 	Chrysler 	as 	noted 	in 	the 	text 	is 	now 	virtually 	in
compliance 	with 	tre 	27.5 	mpg 	standard. 	For 	purposes 	of 	this
analysis, 	we 	assume 	that 	Chrysler 	maintains 	its 	output 	of 	both
small and large cars at 1985 levels in order to continue
complying 	with 	the 	CAFE 	standard. 	Thus, 	FL 	in 	equation 	(AS)
is 	a 	constant 	(= 	FL) •
With 	respect 	to 	U.S. 	imports 	of 	Japanese 	small 	cars, 	we
make 	two 	different 	assumptions, 	as 	described 	in 	the 	text. 	In
simulation 	(1), 	we 	assume 	offshore Japanese 	small 	car 	suppl iers
maintain 	their 	U.S. 	exports 	at 	1985 	levels, 	while 	the 	remaining
members 	of 	the 	non-Chrysler 	fringe 	respond 	to 	small 	car 	price
changes 	with 	changes 	in 	small 	car 	production. 	In 	simulation 	(2),
we 	allow 	offshore 	Japanese 	suppliers 	as 	well 	as 	all 	other 	non-
Chrysler members 	of 	the 	small 	car 	fringe 	to 	alter 	output 	as 	the
small 	car 	price 	changes.
The 	assumed 	specification 	of 	this 	supply 	response 	is of 	the
constant-elasticity 	of 	supply 	variety:
(A9a) 	FS 	= 	Co 	PS 	Cl,
where 	Co 	and 	Cl 	are 	parameters.
6 	AMC 	does 	produce 	one 	large 	car, 	the 	Eagle, 	but 	in 	1985 	only
12,000 	of 	these 	were 	sold. 	Because 	AMC's 	Eagle 	production 	is 	so
small, 	we assume for 	simpl ici ty 	that AMC would simply maintain
Eagle 	sales 	at 	1985 	levels. 	While 	imports 	do 	consist 	of 	large
and 	small 	cars, 	82 	percent 	of 	all 	imports 	fall 	into the 	small-car
class, 	as 	noted 	in 	the 	text. 	Again 	for 	simplicity, 	we 	assume
that 	all 	imports 	are 	small 	cars.
5

Our 	simulations 	assume 	the 	elasticity of 	fringe 	supply 	with
respect 	to 	price 	(CI) 	is 	10.7 	We 	use 	the 	1985 	values 	of 	FS
and 	PS to 	solve 	for 	CO.
With 	these 	fringe 	output 	response 	functions, 	the 	above-
described 	inverse 	demand 	equations, 	and 	equation 	(A6), 	we 	can
transform 	equation 	(AS) 	into 	the 	following:
(A5a) 	R*S 	( 	gO 	(R*S 	+FL)gl
S 	hO 	(S 	+ FS(.))hI)
«S + 	FS(.))g2
«R*S 	+ 	FL)h2
CL 	) 	+
Cs 	) 	= 0
With 	equation 	(A5a) 	and 	constraints 	(A6) 	and 	(A7), 	we 	can 	solve
for S under 	the new CAFE constraint that would be reflected in
the 	value 	of 	R. 	With 	this 	apparatus, 	we 	can 	determine 	the
changes 	in quantity 	and 	price 	that 	would 	result 	from 	increasing
the 	effective 	CAFE 	standard 	to 	27.5. 8
To 	calculate 	the 	post-CAFE 	values 	of 	output 	for 	the 	domestic
producers 	in 	the 	fringe, 	we 	assume 	that 	each 	member 	of 	the 	non-
Chrysler 	fringe 	maintains 	the 	same 	share 	of 	the 	total 	variable
output 	fringe 	production 	that 	they 	had 	in 	1985. 9 	With 	this
assumption, 	we 	can estimate 	the 	effect 	of 	the 	CAFE 	change 	on
domestic 	production 	for 	not 	only 	the 	total 	but 	also 	the 	small 	and
large 	car 	sectors. 	With 	these 	figures, 	we 	can 	use 	labor-output
7 	See 	D. 	Tarr 	and 	M. 	Morkre, 	"Aggregate 	Costs 	to 	the 	United
States 	of 	Tariffs 	and Quotas 	on 	Imports" 	(Federal 	Trade
Commission 	(December 	(1984)) 	at 	63, 	n. 	12.
8 	Equation 	(ASa) 	cannot 	be 	directly 	solved 	for 	S; 	therefore,
once 	the values 	of 	the 	parameters 	are 	determined, 	we 	use 	an
iterative 	search 	procedure 	to 	find 	the 	S-value 	that 	solves 	the
equation.
9 	However, 	recall 	the 	different 	responses 	of 	the 	Japanese
offshore 	small 	car 	suppliers 	in 	simulations 	(1) 	and 	(2).
6
J
I

ratios 	to estimate 	the 	effect 	of 	the 	CAFE 	changes 	on 	automobile
i ndust ry 	employment.
A2. 	The 	Parameters 	Used 	in 	the 	Model.
The starting values for our model for 	the small and large
car 	outputs 	of 	the GM-Ford 	sector 	and 	the 	combined 	domestic-
foreign 	fringe 	sector 	were 	those 	for 	the 	calendar 	year 	1985.
For 	small and large 	car prices, 	we use the 1985 average price for
the 	total 	industry 	for 	each 	type 	of 	ca r. l 0 	These 	initial
quanti ty 	and 	price 	figures 	are 	as 	follows: ll
L = 	3,593,711 	( 1985 	act ual) ,
S = 	3,073,819 	(1985 	act ual) ,
OL 	= 3,965,256 	(1985 	actual),
Os 	= 7,067,082 	(1985 	actual) ,
FL 	= 371,545 	(1985 	act ual) ,
FS 	= 	3,993,263 	(1985 	actual),12
PL 	= $10991,
Ps 	= $9041.
10 	There 	are 	considerable 	differences 	among 	the 	average 	prices
of 	small 	and 	large 	cars 	sold 	by 	the 	various 	suppliers, 	dif-
ferences 	that 	reflect 	different 	model 	mixes 	produced. 	Clas-
sifying cars by a larger number of categories than the two used
here 	would have 	made 	it 	much 	more 	difficult 	if 	not 	impossible 	to
estimate 	the 	social 	loss 	from 	the 	imposition 	of 	CAFE. 	Thus 	we
follow 	recent 	practice 	by 	adopting 	the 	simplifying 	assumption 	of
only 	two 	car 	types, 	large 	and 	small. 	See 	the 	sources 	in 	note 	1
supra.
11 	The 	source 	of 	these 	data 	is 	Automotive 	News 	(January 	14, 	1986)
at 	46. 	The 	1985 	prices 	are 	based 	on 	standard 	models 	without 	options.
12 	As 	previously 	noted, 	all 	imports 	are 	treated 	as 	small 	cars.
In 	1985, 	imports 	totalled 	2,838,116 	units 	according 	to
Automotive 	News 	(January 	14, 	1986) 	at 	46.
7

Because 	we 	assume 	that 	in 1985 	no 	supplier 	was 	earning
excess 	profits, 	Ps 	= Cs 	and 	PL 	= CL' 	In 	addition 	to 	initial
quantities 	and 	prices, 	we 	require 	parameters 	for 	the 	demand 	models.
The 	most 	important 	of 	those 	parameters 	are 	estimates 	of 	the 	price
elasticities 	of 	demand. 	The 	study 	closest 	in 	methodology 	to 	the
present 	work 	suggests 	that 	the 	own-price 	elasticity 	figures 	for
large 	and 	small 	cars 	are 	as 	follows: 13
= -3
= -2.
unfortunately, 	there 	exist 	no 	comparable 	estimates 	for 	the
demand 	cross 	elasticities. 	A large 	number 	of 	econometric 	studies
have 	divided 	the 	automobile 	market 	into segments, 	some with
divisions 	similar 	to our 	distinction 	between 	large 	and 	small
cars, 	and 	none 	have 	found 	a 	statistically 	significant 	coefficient
for 	parameters measuring 	the 	demand 	cross-elasticity 	between 	the
two 	products. 14 	Therefore 	we 	assume 	zero 	cross 	elasticities.
From 	that assumption 	and 	the 	above-mentioned 	own 	price 	elasticity
estimates, 	we 	arrive 	at 	the 	following 	inverse 	demand 	parameters:
gO 	= 	17,396,646.3,
gl 	= -0.333,
13 	See 	J. 	Langenf el d, 	and 	M. 	Munger, 	"The 	Impact 	of 	Fede ral
Automobile 	Regulations 	on 	Auto 	Demand,w 	(Unpublished 	draft,
Fede r al 	Tr ade 	Comm issi on, 	(June 	1985)).
14 	For 	discussions 	of 	this 	issue, 	see 	J. 	Langenfeld, 	"Federal
Automobile 	Regulation," 	ph.D. 	dissertation, 	Washington
University, 	St. 	Louis, 	May 	1983 	and 	J. 	Langenfeld 	and 	M. 	Munger,
"The 	Impact 	of 	Federal 	Automobile 	Regulations 	on 	Autp 	Demand,"
(Unpubl ished 	draft, 	Federal 	Trade 	Commission, 	(June 	1985)).
8

g2 	= 	0.0,
hO 	= 24,034,579.9
hI 	= -0.500,
h2 	= 	0.0.
Finally, 	to 	calculate 	R, 	we 	note 	that 	the 	average 	fleet 	mpg
attained by GM-Ford in MY 1985 was 25.9, while the small 	car mpg
(M s ) 	attained 	by 	GM-Ford 	in 	MY 	1985 	was 	30.103. 15 	Based 	on 	these
two 	figures, 	the 	estimated 	1985 	large 	car 	mpg 	(ML) 	for 	GM-Ford
was 	23.137. 	Using 	these 	figures 	for 	ML and 	MS, 	letting 	K = 	27.5,
and 	substituting 	into 	(A6a) 	provides 	an 	estimate 	of 	R = 	.4585.
15 	~ 	R. 	Heavenrich, 	T. 	Murrell, 	J. 	Chang and 	S. 	Loos,
Light 	Duty 	Automotive 	Fuel 	Economy 	Trends Thru 	1985,
Environmental 	Protection 	Agency 	(1985).
9

Appendix 	II
Deadweight 	Loss 	Calculation
Here 	we 	detail 	a 	method 	of 	measuring 	the 	deadweight 	losses
to 	the 	economy 	imposed 	by 	raising 	the 	CAFE 	standard. 	These
deadweight 	losses 	result 	from 	devoting 	resources 	to 	other 	than
their 	highest-valued 	use, 	and 	this 	kind 	of 	distortion 	can 	be
induced 	by 	the 	imposition 	of 	a 	27.5 	mpg 	CAFE 	standard. 	The
standard 	would 	induce 	an 	increase 	in 	the 	output 	of 	small 	cars 	and
a decrease 	the 	output 	of 	large 	cars from 	what 	it would be in a
free 	market 	equilibrium. 	The 	losses 	would 	consist 	of 	the 	changes
in 	consumer 	and 	producer 	surplus 	that 	result 	from 	the 	adjustments
to 	the 	higher 	standard.
The 	exact 	nature 	of 	the 	CAFE-induced 	changes 	in 	producer 	and
consumer 	surpluses 	in 	the 	large 	car 	market 	differs 	from 	that 	in
the 	small 	car 	sector. 	The 	CAFE 	situation 	in 	the 	large 	car 	market
is 	depicted 	graphically 	in 	Figure 	AI. 	In 	this 	figure, 	price 	is
measured 	on 	the 	vertical 	axis, 	and 	quantity 	is 	represented 	on 	the
horizontal 	axis. 	D represents 	the 	demand 	curve 	for 	large 	cars,
while 	S 	is 	a 	horizontal 	large-car 	supply 	curve 	with 	unit 	average
and 	a 	marginal 	costs 	equal 	to 	CL' 	The 	initial 	equilibrium
position for 	the market 	is at point 	c with price equal 	to CL and
volume 	at 	OLI' 	As 	a 	result 	of 	raising 	the 	CAFE 	standard 	to 	27.5
mpg, 	the 	output 	of 	large 	cars 	is 	reduced 	from 	0Ll 	to 	OL2, 	and
price 	is 	raised 	along 	the 	demand 	curve 	from 	CL 	to 	PL·
The 	deadweight 	loss 	is 	the 	difference 	between the 	gains 	to
producers 	and 	losses 	by 	consumers. 	To 	compute 	this 	difference,
we 	first 	examine 	what 	the 	producer 	lost 	or 	gained. 	With 	the
10

given 	price 	increase, 	producers 	gain 	a 	profit 	equal 	to 	the
difference 	between 	the 	new 	price 	and 	unit 	costs 	(P L - 	CL) 	times
the 	new 	output, 	OL2' 	In 	figure 	Al, 	this 	would 	consist 	of 	area 	A.
The 	change 	in 	consumer 	surplus 	consists 	of 	the 	additional
expenditures 	incurred 	by 	the 	consumers 	of 	large 	cars 	who 	pay 	a
higher 	price 	plus 	the 	loss 	in 	consumer 	surplus 	of 	those 	customers
priced 	out 	of 	the 	market. 	The 	additional 	expenditure 	of
consumers 	who 	continue 	to 	purchase 	large 	cars, 	also 	depicted 	by
area 	A in 	figure 	Al, 	is 	the 	price 	difference 	(P L - 	CL) 	times 	OL2'
The loss by consumers priced out of the market is depicted by the
triangle 	B which 	shows 	the value 	these 	consumers 	would 	have
received 	had 	the 	price 	remained 	at 	CL' 	Total 	consumer 	loss 	then
is 	the 	sum 	of 	areas 	A and 	B (A+B). 	The 	deadweight 	loss 	would 	be
the 	difference 	between 	this 	sum 	and 	the 	producer 	gain, 	area 	A,
which 	happens 	to 	be 	B. 	For 	our 	models 	with 	their 	non-linear
demand 	curves, 	this 	figure 	can 	be 	found 	by 	integrating 	the 	demand
curve 	with 	respect 	to 	price 	over 	the 	distance 	between 	the 	old
price, 	CL and 	the 	new 	price, 	PL' 	That 	integration 	yields 	the
loss 	in 	consumer 	surplus. 	After 	offsetting 	that 	loss 	with 	the
increase 	in 	producer 	profits, 	the 	remaining 	loss 	is 	the
deadwei ght 	loss.
The 	calculations 	of 	consumer 	and 	producer 	surplus 	changes
for 	the 	small 	car market 	is 	somewhat 	more 	complicated 	because 	of
certain 	characteristics 	of 	the 	supply 	side 	of 	the 	market. 	The
situation 	is 	depicted 	in 	figure 	A2. 	The 	demand 	curve 	for 	small
cars 	is 	D. 	On 	the 	supply 	side, 	there 	are 	essentially 	three
sectors; 	the first 	consists of GM and Ford and the second is made
11

up 	of 	the 	firms 	in 	the 	fringe 	that 	do 	not 	alter 	their 	output.
The 	latter 	would 	be 	the 	offshore 	Japanese 	suppliers 	and 	Chrysler
in 	simulation 	(1) 	and 	just 	Chrysler 	in 	simulation 	(2). 	The 	third
sector 	includes 	all 	other 	suppliers. 	GM 	and 	Ford 	are 	assumed 	to
face 	a 	flat 	cost 	curve 	with 	the 	average 	and 	marginal 	costs
equaling 	CS• 	The 	cost 	curves 	of 	the 	second 	sector 	are 	irrelevant
to this analysis because the firms in this sector do not 	change
their 	output 	in 	response 	to 	the 	CAFE 	adjustments. 	For 	the 	third
sector, 	we 	assume 	the 	upward 	sloping 	supply 	curve 	depicted 	by 	F.
At 	the 	original 	equilibrium 	position, 	the 	first 	sector 	and 	second
sector 	firms 	produce 	output, 	Sl' 	and 	the 	third 	sector 	produces 	an
output 	of 	FS I 	where 	its 	marginal 	cost 	equals 	price 	(CS).
Raising 	the 	CAFE 	standard forces 	the 	first 	sector 	firms, 	GM
and 	Ford, 	to 	increase 	small 	car 	sales 	up 	to 	the 	capacity
constraint; 	as 	a 	result 	the 	total 	volume 	for 	the 	first 	and 	second
sectors 	increases 	from 	Sl 	to 	S2. l 	This 	output 	increase 	forces
price 	to 	fall 	along 	the 	demand 	curve 	(D), 	from 	Cs 	to 	Ps 	which
leads to the firms 	in the third sector to decrease their output
along 	fringe 	supply 	curve, 	F, 	from 	FSI 	to 	FS2. 	The 	new 	market
output 	is 02 	(equal 	to S2 	plus 	FS2).
1 	For 	expositional 	ease, 	the 	capacity 	constraint 	is 	not 	depicted
in 	this 	diagram.
12

As 	a 	result 	of 	this 	adjustment, 	the 	producers 	in 	the 	first
and 	second 	sectors 	lose 	an 	amount 	of 	money 	equaling 	the 	price
difference 	(C S - 	PS) 	times 	the 	new 	output, 	s2.2 	This 	would
consist 	of 	the 	areas 	A+B 	in 	figure 	A2. 	The 	firms 	in 	the 	third
sector 	contracting 	output 	along 	their 	supply 	curve 	would 	lose 	an
amount 	of 	money 	equal 	to 	the 	area, 	B. 	If 	one 	moves 	area 	B to 	a
position 	adjacent 	and 	just 	to 	the 	right 	of 	area 	A, 	we 	have 	a
clear 	picture 	of 	the 	total 	producer 	loss; 	it 	is 	equal 	to 	the
area, 	A+B+C+D+E, 	where 	C+D+E=B.
In 	contrast 	to 	producers, 	the 	consumers 	of 	small 	cars 	gain
by the lower price. 	First the people who would have 	bought a car
at the old price gain an amount equal to the difference in price
times the 	old small 	car volume or 	(CS - 	PS) 	times Qli 	it 	is
equal 	to 	area 	A+B+C. 	In 	addition, 	the 	lower 	price 	brings 	new
consumers 	into 	the 	market, 	and 	they 	gain an 	amount 	of 	consumer
surplus 	equal 	to 	area 	D. 	So 	the 	consumers 	gain 	an 	amount 	of
money 	equal 	to 	areas, 	A+B+C+D.
When 	the 	producer 	loss 	is 	subtracted from 	this 	consumer
gain, 	we 	arrive 	at 	the 	total 	deadweight 	loss 	which 	is 	depicted 	by
area 	E in 	figure 	A2. 	For 	our 	calculation, 	however, 	we 	consider
only 	the 	producer 	loss 	that 	born 	by 	American 	firms, 	and 	that
2 	Because 	GM 	and 	Ford 	were 	initially 	earning 	no 	profits 	on 	the
sales 	of 	small 	cars, 	the 	lower 	small-car 	price 	resulting 	from 	the
27.5 	mpg 	standard 	results 	in 	losses 	on 	JM-Ford's 	small 	car
production. 	However, 	because 	the 	higher 	CAFE 	standard 	also
raises 	the 	large-car 	price, 	GM-Ford 	earn 	profits 	on 	large-car
sales. 	In 	order 	for 	GM-Ford 	to 	sell 	large 	cars, 	and 	comply 	with
the 	27.5 	mpg 	standard, 	GM-Ford 	must 	also 	produce 	small 	cars. 	In
effect, 	the 	profits 	from 	large-car 	sales 	of GM-Ford 	subsidize 	the
losses 	from 	small-car 	sales.
13
I!
I

· '
would 	be 	only 	a 	portion 	of 	the 	total 	producer 	loss, 	A+B+C+D+E.
For 	convenience, 	we 	will 	represent 	this 	portion 	by 	the 	term,
M'* (A+B+C+D+E) 	where 	Mis the proportion of the total loss born by
domestic 	firms. 	For 	the 	small 	car 	market, 	then, 	the deadweight
loss 	of 	American 	consumers 	and 	producers 	would 	be 	as 	follows:
Deadweight 	loss 	for 	small 	cars 	= 	(A+B+C+D)- 	M'*(A+B+C+D+E).
Using 	the 	estimates 	developed 	in Appendix 	I, 	we 	can
calculate 	the 	above 	deadweight 	loss 	for 	each 	of 	our 	two
simulations 	by 	combining 	the 	deadweight 	loss 	in 	the 	large 	car
market 	with 	that 	in 	the 	small 	car 	market:
Total 	deadweight 	loss 	= Deadweight 	loss 	for 	small 	cars
+ 	Deadweight 	loss 	for 	large 	cars.
The 	components 	of 	these 	figures, 	consisting 	of 	(1) 	the 	producer
surplus 	change 	in 	both 	the 	large 	and 	small 	car 	markets, 	and 	(2)
the 	consumer 	surplus 	change 	in 	the 	two 	car 	markets 	are 	displayed
in Table 	II 	in the 	text.
14

""	~.
I!l
c:..,
l'n
>~
'"3
::r
l'n
0
t"'"
t"'
I'.J
CI
..,
IJ:)
l'n
(j
Qi
""l	1':)
3:
t"'"
CI
~
""l
';fIi:'
I'D
rt'
o

.'
• •
...,
rn
I\,)
...,
rn
~
...,
....
rn
\Q
~
c::f'1
~
)0
I\,)	
rn
"'i
I\,)
:r
tD
0
rn
~
aDI
~
~
0
I\,)
(')
DI..,
X
DI..,
';1t:"
~
~
o