Cazinouri Canada

Business Profitability versus Social Profitability: Evaluating Industries with Externalities, the
Case of Casinos
Author(s): Earl L. Grinols and David B. Mustard
Source: Managerial and Decision Economics, Vol. 22, No. 1/3, Management and Information
Issues for Industries with Externalities: The Case of Casino Gambling (Jan. - May, 2001), pp.
143-162
Published by: Wiley
Stable URL: http://www.jstor.org/stable/3657264 .
Accessed: 27/03/2014 11:31
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Wiley is collaborating with JSTOR to digitize, preserve and extend access to Managerial and Decision
Economics.
http://www.jstor.org
This content downloaded from 86.55.176.96 on Thu, 27 Mar 2014 11:31:36 AM
All use subject to JSTOR Terms and Conditions
MANAGERIAL AND DECISION ECONOMICS
Manage. Decis. Econ. 22: 143-162 (2001)
DOI:
10.1002/mde.1004
Business
Profitability
v ersus Social
Profitability: Ev aluating
Industries w ith
Externalities,
T h e Case of Casinos
Earl L. Grinolsa,* and Dav id B. Mustardb,*
a
Department of Economics,
Univ ersity of Illinois, USA
b
Department of Economics,
T erry
College
of Business, Univ ersity of Georgia,
USA
Casino
gambling
is a social
issue,
because in addition to th e direct benefits to th ose w h o ow n
and use casinos, positiv e
and
negativ e
externalities are
reaped
and borne
by
th ose w h o do not
gamble.
T o
correctly
assess th e total economic
impact
of
casinos,
one must
distinguish
betw een business
profitability
and social
profitability.
T h is
paper prov ides
th e most
compre-
h ensiv e framew ork for
addressing
th e th eoretical cost-benefit issues of casinos
by grounding
cost-benefit
analysis
on h ouseh old
utility.
It also discusses th e current state of
know ledge
about th e estimates of both th e
positiv e
and
negativ e
externalities
generated by
casinos.
Lastly,
it corrects
many prev alent
errors in th e debate ov er th e economics of casino
gambling. Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
INT RODUCT ION
Betw een 1990 and
1998,
commercial casino rev -
enues increased from $8.7 billion to ov er $22.2
billion,
or 156%.' T h e number of counties w ith
casinos rose from 26 to almost 200 in th e same
time.
Including
Class III American Indian casi-
nos,2 casino rev enues totaled $29.5 billion in 1998,
representing expenditures
of $153
per
adult
aged
20 or ov er.
T h e
rapid expansion
of casinos to new
parts
of
th e
country generated
extensiv e debates about th e
impact
of casinos on a
range
of
social, economic,
and
political
issues.3 T h ese concerns w ere suffi-
ciently pronounced
to cause
Congress
to establish
th e National
Gambling Impact Study
Commis-
sion
(NGISC)
in 1996 to conduct an exh austiv e
study
of th e
impact
of casinos.4 At th e conclusion
of its
inv estigation,
th e commission recommended
a national moratorium on th e
expansion
of
gam-
*
Correspondence
to:
aDepartment
of Economics,
Univ ersity
of
Illinois,
1206 S. 6th
Street, Ch ampaign,
Illinois 61820.
E-mail:
grinolsGuiuc.edu; h Department
of Economics,
T erry
College
of
Business, Univ ersity
of
Georgia,
528 Brooks Hall,
Ath ens, Georgia
30602. E-mail:
mustard@terry.uga.edu
bling
and more
study
of
gambling's
effects,
costs
and
benefits,
before
making
furth er decisions
about it.
T h e literature on th e costs and benefits of
casino
gambling
is
fraugh t
w ith
inadequacy
and
confusion. Ev en studies th at
purport
to ev aluate
th e economic
impact
of casinos
commonly
exh ibit
a
great
deal of
misunderstanding
about w h at
sh ould be included
among
benefits and
costs,
and
prov ide
little or no
guidance
about h ow th e costs
and benefits relate to one anoth er or sh ould be
computed. Many
studies
pay
a
great
deal of atten-
tion,
for
example,
to
estimating
th e number of
direct and indirect
jobs
th at casinos create and to
tallying
th e taxes casinos
pay,
but do not
explain
th e social v alue of an additional
job
or calculate
th e lost taxes of
competing
non-casino busi-
nesses.5 In
general,
th e costs and benefits dis-
cussed are
casually
listed,
v ary by study,
and are
commonly presented
w ith little or no
justification
of h ow
th ey
w ere selected or
w h y
oth er
potential
costs and benefits w ere excluded.
A recent
paper, Eadington (1999),
is instructiv e.
It identified th ree
principal
benefits of casinos:
(1)
Copyrigh t
? 2001 Joh n
Wiley
& Sons, Ltd.
E.L. GRINOLS AND D.B. MUST ARD
gain
in
utility
(for th ose
gambling
in moderation
for
entertainment), (2) ancillary
economic benefits
such as
'job creation,
inv estment
stimulation,
tourism
dev elopment,
economic
dev elopment
or
redev elopment,
urban or w aterfront
rev italization,
or th e
improv ement
of th e economic status of
deserv ing
or
underpriv ileged groups',
and
(3)
ad-
ditional rev enues to th e
public
sector. He lists tw o
principal
costs:
(1)
'moral
disapprov al'
and
(2)
'fears of adv erse social
impacts',
such as
path olog-
ical
gambling,
crime, or
political corruption.
T h e
net increase in
profits
to
business,
unless th is is
meant to be
part
of
ancillary
economic
benefits,
is
absent from th e list of benefits.6
Alth ough
Ead-
ington
lists
gain
in
utility (clearly
internal to th e
indiv idual or
h ouseh old)
as a
benefit,
h e w rites
th at
'many
of th e costs identified are internal to
th e indiv idual or th e h ouseh old, as
opposed
to
external-borne
by society-and
are th erefore
difficult to
place
into a
cost/benefit
framew ork'.
T h is v iew of costs
(including
th e references to
moral
disapprov al
and fears of
consequences
in-
stead of th e actual
consequences) suggests
th at th e
auth or believ es costs are more subtle and
possibly
less
tangible
th an benefits.
How ev er, because th e
process
to determine h ow items are included is
not
explained,
th ere is little th eoretical
guidance
about h ow th e identified cost-benefit
components
relate to one anoth er in an ov erall assessment of
th e
impact
of casinos or h ow
competing
costs and
benefits are reconciled. We w ill sh ow h ow cost-
benefit
components
based on
utility
can be
placed
into th e ev aluation framew ork.
T o
bring uniformity
and more
th eory
to bear
on th e cost-benefit treatment of
casinos,
th is
paper
demonstrates th e construction of an ex-
h austiv e and
utility-grounded
framew ork to iden-
tify
costs and benefits. It outlines an
explicit
taxonomy
for costs and benefits based on th e
principle
of real resource
use,
and rev iew s th e
av ailable studies th at contain
original
research
estimating
one or more cost-benefit
components.
Alth ough
th e
primary purpose
of th is
paper
is to
rectify
th eoretical cost-benefit
reasoning
as it
ap-
plies
to
casinos,
th e
meth odology applies
more
generally
to th e ev aluation of
projects
in oth er
industries. We also rev iew
existing empirical
esti-
mates of th e costs and benefits of casinos ar-
ranged according
to th e
th eoretically grounded
principles. Unfortunately,
th ere h as been rela-
tiv ely
little research on
many
of th e most
impor-
tant social cost-benefit
components,
w h ile much
of research h as examined less
significant
issues or
issues th at are not ev en
part
of a
properly
defined
analysis
of social costs and benefits. Some re-
search th at
purports
to ev aluate costs or benefits
actually
examines local and not total social costs
or benefits. Anoth er concern is th at much of th e
research h as been conducted
by organizations
w ith a v ested interest in th e outcome of th e re-
search ,
institutes w ith
industry ties,
or state
agen-
cies.
Relativ ely
little research is in
peer-rev iew ed
journals.
A rev iew of th e
empirical
literature th at
estimates
correctly
defined
components
of social
costs and benefits indicates th at th e costs of casi-
nos are at least 1.9 times
greater
th an benefits.
T h e remainder of th e
paper
is
arranged
as
follow s. T h e next section constructs a th eoretical
cost-benefit measure based on economic funda-
mentals. T h e th ird and fourth sections examine
th e social benefits and costs of casino
gambling,
respectiv ely.
T h e fifth section concludes
by
sum-
marizing
our contributions and
outlining
th e im-
plications
of th is w ork for future research .
T HEORY
Linking
Cost-Benefit to
Utility
In th is
section,
w e
lay
out th e foundations of
cost-benefit
analysis
for casino
gambling.
T o
av oid th e mistakes th at h av e
plagued
cost-benefit
analyses, especially
confusion about w h at can be
included on each side of th e cost-benefit
ledger
and h ow each item sh ould be
computed,
w e start
from th e most fundamental cost-benefit
concept
possible-indiv idual utility.
T h e framew ork w e
employ
can be as
compreh ensiv e
and
general
as
desired, alth ough
our
objectiv e
is to
prov ide just
enough
detail to include all of th e
major
elements
commonly
considered relev ant to th e economic
effects of
gambling
and
enough explanation
to
indicate w h at w ould
ch ange
in a more detailed
application
of th e framew ork.
Our
starting point
is th e
ch ange
in th e indiv idu-
al's
utility,
u -
u?,
w h ere
superscripts distinguish
utility
in tw o situations. In one, casinos are w ide-
spread geograph ically
(alternativ e 1)
and in th e
oth er,
casinos are less
w idely spread
(alternativ e
0).
We assume th at
u(x)
is a continuous
utility
function
representing locally
non-satiable
prefer-
ences defined on
consumption
xeR". A
positiv e
element of x denotes
consumption
of a
good
or
serv ice,
w h ile a
negativ e component
stands for th e
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
144
Manage.
Decis. Econ. 22: 143-162
(2001)
BUSINESS PROFIT ABILIT Y VERSUS SOCIAL PROFIT ABILIT Y
prov ision
of a
good
or serv ice.7 For
example,
th e
prov ision
of 10 h ours of labor
by
th e indiv idual
w ould
appear
as -10 in th e labor
component
of
x. We define th e
expenditure
function
e(d, p. u)
as
th e minimum
expenditure
needed to ach iev e util-
ity
u w h en th e consumer
buys
and sells at
prices p,
and d is th e distance to th e nearest casino. It is
strictly
monotonic in u for
any
ch oice of fixed d
and
p.
T h e
sign
of
e(d', p',
u')-e(d, p', u?) is,
th erefore,
identical to th e
sign
of u'
-
u?. In oth er
w ords,
for fixed distance and
prices
d and
p, e(d,
p, u(x))
is a
utility
function w h ose natural
money
metric records
utility
in dollars.8
We
compare
th e social w elfare betw een th e tw o
situations. We
presume
for
simplicity
th at
gam-
bling
is a standardized
good;
casinos offer
gam-
bling
on
essentially
th e same terms as casinos in
oth er locations.9 T h e
primary adv antage
to th e
consumer of more
casinos, th erefore,
is closer
proximity
to th e nearest one. Let d' be th e dis-
tance to th e nearest casino for consumer i in th e
post
casino alternativ e 1. Our measure of social
profitability
is th e
ch ange
in w elfare for all
consumers
AW=
Z w ,[e,(dl,
p,
u,)
-
e1(d',
p ', u?)],
w h ere
j w i
=
m,
and m is th e number of con-
sumer h ouseh olds.
Equation (1)
allow s for differ-
ent
w eigh ts
for dollar
gains
to different
h ouseh olds,
a
topic
to w h ich w e w ill return below .
How ev er,
in
applying (1)
to
produce
a
w orking
measure of social
profitability,
w e
explicitly
ad-
dress
many
issues left
unspoken
in some studies
and th at are a source of confusion in oth ers. T h e
initial model
prov ides
th e
simplest
framew ork for
analyzing
th e
impact
of casinos. We list our as-
sumptions
at th e outset for
clarity.
-
We assume th at a dollar of
utility
to one
h ouseh old is
equal
to a dollar of
utility
to
anoth er.'? With
respect
to
Equation (1)
th is
implies
th at w
=
1 for all h ouseh olds. It also
means th at firm
profits
do not need to be
assigned
artificial
premia
or discounts based
on w h ich indiv iduals or h ouseh olds
h appen
to
ow n th em.
-
Firm
profits
are
equally important
to social
w elfare
regardless
of w h ich firm
generates
th em. For
example,
casino
profits
are v alued
th e same as th e
profits
of a non-casino firm.
-T o allow for
regional
tax differences, con-
sumers and firms
may
face different
prices.
In
th e
limit,
each firm and h ouseh old could h av e
a different,
personalized
set of
prices.
House-
h old i faces
price
v ector
pi,
firm
j
faces
price
v ector
p,
and endow ments are traded at
prices
Pno
-We allow for th e
possibility
th at consumers
may
be constrained in th eir labor
supply
deci-
sions,
resulting
in
unemployment. People
h av e
a reserv ation
w age
but cannot
alw ays
find a
job
at th at
w age,
and
low ering
th eir
w age
w ill
not increase th e ch ances of th eir
getting
a
job.
-
Firms and
economy
endow ments are ow ned
by
h ouseh olds. Househ old i ow ns sh are
0,0
of firm
j, E,0,=
1 and endow ment
QeR'+,
w h ere
Si
f, =
Q,
th e
economy
endow ment v ector.
-
T h e
gov ernment spends
tax rev enues to
pur-
ch ase
goods
and
serv ices,
and
priv ate
h ouse-
h olds receiv e
utility
from th ese
expenditures.
T o
implement
th is
assumption,
w e
employ
th e
artificial dev ice of
h av ing
th e
gov ernment
re-
turn tax dollars to h ouseh olds in a
lump-sum
fash ion. Househ olds th en
spend
th e transfers
as
part
of th eir income and
experience utility
gains
based on th eir
purch ases.
-
In addition to direct benefits and
costs,
casinos
may generate positiv e
or
negativ e
externalities.
Positiv e externalities add v alue to th e
economy
not
reaped by
th e
agent creating th em,
w h ile
negativ e
externalities remov e v alue not
paid by
th e
causing agent, follow ing
th e usual defini-
tion. For
example,
if a casino's
presence
re-
duces crime in an
area, leading
to less need for
police presence,
th is frees real resources to th e
rest of th e
community
and
represents
a
posi-
tiv e
externality.
If th e rev erse is
true,
and th e
casino increases th e need for
police,
real re-
sources are remov ed from final
consumption
x, and th is is a
negativ e externality.
T h e th ird
and fourth sections discuss th e nature of bene-
fits and costs in more detail. T h e net resources
gained
or lost to th e
system
are denoted
by g.
If
g
> 0
negativ e
externalities
outw eigh positiv e
externalities,
w h ich decrease th e resources
av ailable for
consumption
x,
and
th ereby
low er social w elfare. Social cost
accounting
in
real terms
requires
x + g
=
y + Q
+,
w h ere x =- ,
xi
is
aggregate consumption,
and
y
-
?
y
is
aggregate production.
For each
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
145
Manage.
Decis. Econ. 22: 143-162
(2001)
firm
j, yj
is th e associated
production
v ector";
z is th e
economy
trade v ector.'2
T h e abov e remarks
prov ide
th e
simplest
frame-
w ork th at is
sufficiently
inclusiv e to discuss an
economy's
social costs and benefits of
gambling.
Application
Consider now th e
follow ing carefully
ch osen iden-
tity,
a
telescoping
sum w h ere each term cancels
part
of th e
preceding
term.
[ei(d', p', u!)- e,t(d, p, u)]
=
[e,(d', p', u')-p 'xI] (2.1)
(Consumption
Constraints in Situation
1)
+
[p'x -p?.x,]
(2.2)
i
(Income Effects)
+ E [pO.x?
-
e,(dOp p, u')] (2.3)
i
(Consumption
Constraints in Situation
0)
+
E
[e,(d?,
p?,
u?)
-
e,(dl, p?, uO)] (2.4)
(Distance
Benefits)
+
E [ei(d, po, uo) -
e,(d,
pp,
u
)] (2.5)
(Consumer Surplus)
Expression (2.1)
measures th e w elfare
impact
of
constraints on th e consumer's ch oice th at
prev ent
h im from
being
at h is
optimal
bundle
giv en
th e
prices
h e faces. T h e
primary example
of th is kind
of constraint is
unemployment. e,(di,
p,, ul) by
definition is th e least
costly w ay
of
ach iev ing
th e
utility
ach iev ed in situation 1.
Consumption
bun-
dle x satisfies u'
=
u(xJ)
and also ach iev es
utility
u'. Because ch oice of
x]
w as constrained (in th e
case of
unemployment, by
th e consumer's
ability
to
supply labor),
it w ill lead to a
greater expendi-
ture th an
e,(d], p , u'). T h erefore,
th e difference
in
expression (2.1)
is th e amount of
money
th e
indiv idual w ould be
w illing
to
pay
to remov e th e
constraint. T h e same
argument applies
to
expres-
sion
(2.3)
in situation 0.
Expression (2.4)
measures th e v alue to th e con-
sumer of
h av ing
th e nearest casino distance d'
aw ay compared
to distance d?. For
example,
in
th e initial situation th e consumer needed
e,(d',
p',
u")
to reach initial
utility.
Wh en th e nearest casino
is
closer, distance
dj
<
d?,
th e income needed to
maintain
original utility, e,(d, p?, u?),
is smaller
(presuming
th e indiv idual
gambles).
T h e differ-
ence in
expression (2.4), th erefore,
is th e amount
th e consumer w ould be
w illing
to
pay
to h av e th e
nearest casino closer.
Expression (2.5)
is th e conv entional measure of
consumer
surplus.
T h e
only
difference betw een
th e tw o terms in th e
expression
is th e
price
v ector.
If
prices p)
are better for th e h ouseh old th an
prices p? (low er
for
goods purch ased and/or
h igh er
for
goods sold,
such as
labor),
th en
expres-
sion
(2.5)
is
positiv e
and measures th e amount of
money
th e consumer w ould be
w illing
to
giv e up
to h av e th e better set of
prices.
Now examine
expression (2.2).
Use th e h ouse-
h old
budget identity
Pi xi = E o,inj
+
n
*
ni
+
T
-
E
J
(3)
to transform th e income effects in
(2.2)
w h ere Il
is th e
profit
of firm
j, pn
iQ
is
earning
from th e
h ouseh old's
endow ment, T i
is th e h ouseh old's
sh are of
taxes,
and
E,
is th e h ouseh old's sh are of
th e cost of
gambling-induced externality expendi-
tures.
Summing (3)
ov er h ouseh olds and differ-
encing
betw een th e initial and final situations'3
yields
Z [P .X' -p? x0 = A1ij (AProfits)
[i j
+
Apn
Q
(Endow ment
Capital Gains)
+ AT
(AT axes)-
AE
(AExternality Costs) (4)
Substituting (4)
into
(2); w riting
th e distance, ef-
fects in differential form and
rearranging giv es
th e
taxonomy
of cost-benefit elements th at w e seek:
[e,(d), p', u) -
e(d',p',
u?)] AW=
E A+Xdd +eAT -AE
Consumption Constradd ints +
-
AE
J
i
Jdonip.
+
Consumption
Constraints+
Ap0o
Q
+
[e,(d', p?, uo)
-
e,(dl,
p1, u')],
/
(5)
w h ere
'Consumption
Constraints' is th e sum
(2.1)
+
(2.3).
T h e sev en
components
in
Equation (5)
are an
exh austiv e,
exact tabulation of th e cost-benefit
Copyrigh t
? 2001 Joh n
Wiley
& Sons, Ltd.
146
Manage.
Decis. Econ. 22: 143-162
(2001)
elements for
ev aluating
th e economic effects of
casinos. Moreov er
(5)
sh ow s
precisely
h ow each
term sh ould be
computed th eoretically.
For exam-
ple,
th e effect of casino
gambling
on firm
profits
sh ould be summed ov er all
firms,
not
just
casinos.
T h e increased
profits
of th e casinos sh ould be
netted
against
lost
profits
of oth er firms th at
compete
for consumer
spending. Comparable
statements
apply
to th e
computation
of
employ-
ment benefits and
costs, taxes, and social costs.
T h ere is one obv ious
simplification
w e can
make to
(5).
Because
gambling industry
rev enue
(casinos, lotteries,
racetracks and oth er forms of
gambling)
is
relativ ely small,'4
it w ill h av e a
negli-
gible
effect on
creating capital gains
or losses on
endow ments. It is
unlikely
th at th e cost of
capital,
for
example,
w ill differ because of th e
presence
or
absence of casinos in th e
economy.
A similar
statement
applies
to consumer
surplus
effects th at
depend
on
gambling
to influence ov erall
prices.'5
T h erefore, for th e remainder of th e
paper
w e
assume th at firm and h ouseh old
prices
are inv ari-
ant to th e amount of
gambling (p?=p], p
=
p
,.
PO
=
pn),
w h ich means th at th e last tw o terms
in
Equation (5)
related to
capital gains
on endow -
ments and consumer
surplus gains
and losses
drop
out.
Conceptual
Corrections
Equation (5)
allow s us to address some common
errors and
misconceptions
of cost-benefit
analy-
sis
applied
to
gambling.
T h e first error is th e
tendency
to
identify
busi-
ness
profitability,
Ej
HI,
and its
improv ement,
iE
All,
w ith social
profitability.
T h e tw o are dif-
ferent. Business
profitability
is
clearly important
to social
profitability
and contributes to it, but th e
tw o are not
synonymous.
Failure to account for
all of th e
components
of social
profitability
is
perh aps
th e most common mistake. Casino
profits
are v isible and
prominent.
Oth er costs and bene-
fits
may
be less so.
T h e second error is to ev aluate th e economic
impact
of
gambling
w ith
respect
to th e taxes and
profits
of a subset of
firms-typically
th e
profits
of firms in one state or
region
and sometimes th e
profits
of local
gambling
firms
only. Equation (5)
sums
profits
ov er all
firms, not
just
casinos or
firms in one location.
Ignoring
firms th at lose
profits
due to th e
expansion
of
gambling
is
equiv -
alent to
selecting w eigh ts
for th em in
Equation (1)
th at are zero. Because h ouseh olds ow n th ese oth er
firms, th is v iolates th e
assumption
th at h ouse-
h olds are treated
equally.
T h e th ird is to consider
only
th e taxes of a
subset of h ouseh olds or
regions.
It is not uncom-
mon,
for
example,
for studies to focus
only
on
costs w ith in th e
state, ev en
th ough
casinos th at
border anoth er state h av e ramifications for citi-
zens of th e
neigh boring jurisdiction. Equation (5)
sums taxes ov er all h ouseh olds and
regions.
Ev aluations th at consider
only
th e costs or
benefits of a subset of h ouseh olds or
regions
are
inaccurate and
incomplete.
For
example,
th e
cost-benefit measure in
(5)
does not treat a
job
in
a
giv en
location as more v alued th an a
job
in
anoth er location.
Many
economic
impact
studies
perform regional
net
export multiplier analyses
of
th e effects of casinos.
T h ey erroneously report
th e
number of
jobs
in a
giv en
location as a benefit.
According
to
(5)
th e v alue of
employment
in one
location
(part
of th e determination of firm
profits)
must be netted
against
th e v alue of
employment
in
anoth er location. T h ere is no net
gain
to th e
economy
from
sh ifting
a
job
from one location to
anoth er unless it increases
profits
to th e
economy.'6
T h e last common error is th at much
empirical
w ork
purports
to sh ow casinos decrease unem-
ployment,
but fails to
prov e
w h at
employment
w ould h av e been in th e absence of casinos. Most
casinos w ere introduced after
1991,
w h en th e
country
w as
recov ering
from th e recession of
1990-1991. T h e
period
from 1991 to 2000 also
coincided w ith th e
longest
economic
expansion
in
American
h istory.
As th e
country emerged
from
th e
recession,
unemployment
declined in areas
w ith and w ith out casinos. If casinos
temporarily
reduced
unemployment
faster th an it w ould h av e
fallen
oth erw ise,
th is
transitory
effect could cor-
rectly
be counted as a benefit of casinos. How -
ev er,
w e know of no
study
th at h as made th is
case. On th e
contrary,
th e failure to account for
th e decline in
unemployment
th at w ould h av e
occurred
anyw ay
leads to a classic
post
h oc,
ergo
propter
h oc
fallacy
of
logic.
For a more detailed
example,
see
Appendix
A,
w h ich discusses T h e
Ev ans
Group (1996). Alth ough
it
argued
th at
casinos reduced
unemployment,
it did not
report
th at areas w ith out casinos w ith
comparable
start-
ing unemployment
rates
experienced comparable,
and in
many
cases,
larger
reductions in th e unem-
ployment
rate.
Copyrigh t
C 2001 Joh n
Wiley
&
Sons, Ltd.
147
Manage.
Decis. Econ. 22: 143-162
(2001)
BOUNDING BENEFIT S
T h is section rev iew s th e studies th at estimate th e
benefits from casinos based on th e
th eoretically
correct cost-benefit
computation
in
Equation (5).
We discuss in order th e net increase in firms'
profits plus
taxes
paid
due to th e
presence
of
casinos,
th e consumer distance benefits of nearer
casinos,
employment
benefits and total benefits
from th e
expanded gambling opportunities.
Profits and T axes
T h is benefit is calculated
by determining
th e
casino
profits
and taxes minus th e reduction in
profits
and taxes of oth er businesses caused
by
casinos.
Alth ough
casino
profits
and taxes are
h igh ly v isible, th ey
are inv alid measures of social
benefits because
th ey
do not
adjust
for th e entire
economy
for th e lost
profits
and taxes of
compet-
ing
businesses. T h is
point
is not
special
to casinos.
Any
business-be it Wal-Mart or a
drugstore
ch ain,
th at attracts consumer
sales, employs
labor
and oth er
inputs,
and
displaces competing
busi-
nesses-sh ould be ev aluated on th e same basis.
Because
many
casinos do not h av e to
report
th eir
profits
or
pay
taxes
(for
example,
casinos
ow ned
by
American Indian
tribes),
th ere are no
data on
industry profits. How ev er, w e can esti-
mate rev enues from
annually publish ed
informa-
tion. We
prov ide
a brief ov erv iew of casino
gambling
in th e US before
estimating
th e benefits.
T able 1
reports
total and
per capita gambling
rev enue.17 For
comparison,
w e
prov ide
data on
th e tobacco
industry.'8 Many
studies estimate
po-
tential casino rev enues
using
th e amount of
gam-
bling per person
in areas w h ere casino
gambling
is
a
prominent activ ity.
For
example,
th e
City
of
Ch icago Gaming
Commission funded a
study
(Deloitte and
T ouch e, 1992)
th at
reported
th at
adults w ith in 35 miles of Atlantic
City
lost $198
per
adult
annually
to casinos.
Adjusted
for
eigh t
years
of
price ch anges,
th is
figure
is
approxi-
mately
$230. In its
study,
th e
Mirage
Hotel
(1993)
estimated th at annual
per capita gambling
rev -
enues for
persons residing
w ith in a 50-mile radius
of its
proposed Ch icago
suburb
gambling facility
w ould be $200.'9 In Iow a, in 1995, a Ch ristiansen
and
Cummings
Associates
study
for th e state
Racing
and
Gaming
Commission found th at th e
av erage
adult lost $172 to th e casinos (th is figure
is low er th an $230 because casinos are still not in
close
proximity
to all
parts
of
Iow a).
T h ese data
are
comparable
to rev enue for oth er areas.
In addition to
av erages
w e are interested in th e
concentration of
gambling among
users.
Many
studies examined
gambling
markets in different
locations and at different times. T aken
togeth er,
th ey prov ide
a
general
estimate of h ow
frequently
residents
gamble.
In a market w ith
readily
av ail-
able
gambling opportunities including
casinos, ap-
proximately
30'% of th e
population
does not
gamble, meaning
th at
th ey
w ill not h av e
gambled
in th e
past year.20
Anoth er 50-60% could be
termed
ligh t bettors,
w h o
gamble
less th an once
per
w eek. T h is
group
includes th ose w h o
enjoy
a
nigh t
out at th e casino once in a
w h ile,
but do not
frequent
casinos. About 5-15% could be termed
h eav y
bettors w h o
gamble
tw ice
per
w eek or
more. T h e last 2-5% of th e
population
consists of
problem
and
path ological (P&P) gamblers,
w h o
suffer from
compulsiv e gambling disorders,
w h ich
are
expressed
w h en th e
opportunity
to
gamble
is
present
and sufficient time h as
elapsed
for th e
problem
to become ev ident. T h is
group migh t
be
in th e casino
daily,
for
long periods
of
time,
and
at unusual h ours. T w o-th irds to 80% of
gambling
rev enues come from th e 10% of th e
population
th at
gambles
most
h eav ily.21 Expressed
in
rev erse,
90% of th e
population may prov ide
as little as
20%/o of casino rev enues.
Consequently,
th e
great
majority
of adults are
indifferent,
or
nearly
indif-
ferent, to th e
av ailability
of casino
gambling.
Alth ough
th e
av erage
American adult loses
T able 1. T h e Casino Market
Casino
gambling
rev enues All
gambling
rev enues T otal rev enues
($)
All
gambling
per
adult
($) per adult ($)
rev enues
($)
US 1998 153 282 29.5 billion 54.4 billion
'Saturated market' _ 230 359 44.4 billion 69.3 billion
T obacco
industry 39 billion
Copyrigh t
(
2001 Joh n
Wiley
&
Sons, Ltd.
148
Manage.
Decis. Econ. 22: 143-162
(2001)
approximately
$153
per year
and
migh t
lose closer
to $230
per year
w ere
gambling
more
w idespread,
th ese rev enues come from a few w h o
gamble
a
lot,
instead of
many
w h o
gamble
a little.
We now return to our
original question
w h at
is th e social v alue of th is amount of casino
gam-
bling? According
to
Equation (5)
w e need th e
profits
and taxes attributable to
casinos, minus
th e reduction in
profits
and taxes of oth er busi-
ness due to casinos. T o th ese w e must add th e
consumer distance benefits of casinos
(w h ich
w e
address in th e next
section).
Because
profits
are a
function of market structure and th e
presence
of
free
entry
and
exit,
if casinos w ere
deregulated,
market
contestability
and free
entry
of casinos
w ould driv e economic
profits
to zero. In th at
ev ent,
from th e
perspectiv e
of
profits,
a
larger
casino sector and smaller remainder of th e econ-
omy
w ould
represent
a net w ash because eco-
nomic
profits
in th e
economy
w ould be no
greater
w ith casinos th an w ith out. T h e sole contribution
of casinos to social w elfare in th at case w ould be
th e direct consumer benefits.
How ev er,
in th e current
legal env ironment,
casinos in
many
locations are
effectiv ely regional
monopolies
sustained
by gov ernment licensing
re-
strictions.22
We, th erefore, make th e
follow ing
adjustment
to allow for th e
h igh er monopoly
profits
of some casinos. In
1998,
profits
before
taxes23 of all non-financial
corporate
business in
th e United States w ere 13.8% of sales.24
Assuming
th at casinos
av erage
30%
profit
rates before taxes
(more
th an double th e normal business rate of
profit) implies
th at social benefits in th e form of
profits
and taxes from
sh ifting
$153 of rev enue
from oth er businesses to casinos is
(0.30-0.138)
153
=
$25 rounded
up
to th e nearest dollar. In th e
next
section,
w e add to th is consumer distance
benefits of casinos to
produce
an
upper
bound on
total casino social benefits.
Consumer Distance Benefits
Equation (5)
also identifies
Jd (Sei/d,)dd,
as a
direct social benefit of
casinos,
w h ere
d,
is con-
sumer i's distance to th e nearest casino. Distance
benefits h av e been little studied,
ev en
th ough th ey
constitute a
primary
direct benefit of casinos. T o
our
know ledge, only
Grinols
(1999)
estimated
th ese benefits and
compared
th em w ith th e oth er
components
of
(5). Assuming
th at
utility depends
on
goods
x,
th e number of casino v isits V. th e
amount
gambled (spent) per
v isit
g,
and th e dis-
tance trav eled to th e
casino, u
=
u(g, V, I(g, d))
w h ere
I(g, d)
is an
enjoyment
factor or v isit 'in-
tensity'
factor th at rises w ith
g
and falls w ith d
and
g
is
consumption
of oth er
non-casino, goods.
T h e
env elope
th eorem and consumer
optimization
conditions sh ow th at
Jdp
(ae,/d,)dd,
< Vdg.
T h is
inequality
allow s inferences about w elfare to
be made from data th at relate to th e number of
v isits and amount
gambled
per
v isit to th e dis-
tance from th e casino. Grinols
(1999)
estimated
th at th e
upper
bound for direct conumer benefits
of casinos w as $50
per
adult
(again, rounding up
to th e nearest round
figure
to
produce
an
upper
bound on casino
benefits)
w h en no allow ance is
made for th e
significant portion
of rev enues from
problem
and
path ological gamblers.
If th e rev -
enues of non-P&P
gamblers only
are used to
calculate consumer distance
benefits,
th en th e
benefit
figure
falls to under $34.25 T h is number
can be
interpreted
as th e answ er to th e
question,
'How much w ould
you
be
w illing
to
pay
each
year
to h av e th e
opportunity
to
gamble
in a casino
nearby compared
w ith th e alternativ e w h ere casi-
nos are 1000 miles
aw ay?'
Employment
Benefits
Alth ough
th e
topic
of
employment
benefits is one
of th e most studied issues about casino
gam-
bling,26
it also contains a
w idespread
and central
misunderstanding-th at
th e benefits of new busi-
nesses are measured
by
th e
jobs th ey
create in a
giv en
location. Wh ile it
may
be
legitimate
to ask
w h at effect a new business w ill h av e on
employ-
ment,
w h at taxes it w ill
pay,
and from w h ere its
rev enues w ill
come,
th ese answ ers do not assess
th e social benefits and costs of th e business. In-
creasing jobs
in one location at th e
expense
of lost
jobs
in anoth er is not a social benefit. Business
profitability
is not social
profitability.
Social cost
benefit is
grounded
on consumer
utility
and re-
sults in a list of relev ant factors different from
tracking
income and
employment
effects.:7
T otal Social Benefits
Based on th e
prev ious
sections,
if casinos w ere
fully deregulated
and allow ed to
spread freely
nationw ide,
economic
profits
w ould be driv en to
zero. T h e net increase in
profits
and taxes from
Copyrigh t
? 2001 Joh n
Wiley
& Sons, Ltd.
149
Maonage.
Decis. Econ. 22: 143-162
(2001)
expanding
th e casino sector at th e
expense
of th e
rest of th e
economy,
th erefore, w ould be zero.
T h e consumer distance benefits of casinos w ould
be less th an $50
per adult,
or if th e rev enues of
P&P
gamblers
are
subtracted, $34
per
adult.
If casinos are
regulated
and
granted regional
monopoly
status in some
jurisdictions,
th e eco-
nomic
profits
of casinos w ill remain
positiv e,
but
th e distance benefits w ill
drop. Assuming av erage
pre-tax profits equal
to 30% of sales
(more
th an
double th e rate for non-financial
corporate
busi-
ness in th e
US) implies
th at th e net
profit
and tax
benefits of casinos are less th an $25
per
adult.
How ev er,
if th ere is not free
entry,
distance bene-
fits w ill
av erage
less th an $50
per
adult
(less
th an
$34
adjusting
for P&P
gamblers)
because some
areas w ill not h av e casinos close to consumers.
We are, th erefore,
left w ith th ree
upper
bounds.
T h e
preferred number, $34
per
adult,
is th e most
correct
upper
bound because it
represents
th e full
social v alue of casinos under circumstances in
w h ich all of th e benefits w ould be
captured by
consumers if th e
industry
w ere
deregulated
to
allow free
entry. Fifty-nine
dollars combines th e
full estimate of consumer distance benefits ad-
justed
for P&P
gamblers
w ith a
generous profit
figure.
It is too
h igh
because th e consumer benefit
is ov erstated, and in
addition,
because it fails to
recognize
th at distance benefits w ould decline
w ith
regional monopolies present
th at do not
put
casinos close to all consumers.
Finally,
$75
per
adult adds consumer benefits to
profits
w ith out
making any adjustments.
We
emph asize
th at th ese
numbers are
upper
bounds on th e estimated
benefits.
COUNT ING COST S
Research ers estimate th e social costs of casinos
using
tw o meth ods. T h e first is
th rough
th e
study
of
problem
and
T h e sec-
ond is
th rough
statistical
analyses
of
cost-creating
activ ities such as
crime, suicide,
and
bankruptcy.
T h e former
approach
ties th e cost activ ities to
gamblers,
but ov erlooks social costs th at do not
deriv e from
problem
and
T h e latter
approach , determining
th e effect of
casinos on social costs such as crime
by examining
aggregate
statistics, is direct and more inclusiv e
because it looks at more th an
just
th e crimes
committed
by
P&P
gamblers.
T h e remainder of th is section consists of tw o
parts.
T h e first deriv es a detailed
taxonomy
of
cost classifications tied to th e th eoretical
analysis
in th e second section. Wh en
discussing
th ese clas-
sifications, w e cite cost studies of both
types
listed
abov e. T h e second
part
of th is section is a more
detailed rev iew of all th e studies th at focused
specifically
on
problem
and
path ological gam-
blers. We calculate costs
per path ological
and
problem gamblers,
and estimate th e costs for th e
entire nation. T h ese sections constitute th e most
compreh ensiv e compilation
of th e social costs of
gambling
av ailable to date.
Cost
T axonomy
T h e
underlying principle,
based on
Equation (5),
is th at each social cost uses
ph ysical
resources
g
in
w ays
th at do not
directly
enter
utility
or th at
reduce economic
efficiency.
We
arrange
social
costs into nine
disjoint groups
and discuss each
one
briefly.
1. Crime: Of all th e social costs,
th e link betw een
casinos and crime h as receiv ed th e most atten-
tion.28 Crime costs are real resources used for
th e
appreh ension, adjudication,
incarceration,
and reh abilitation of
criminals,
or th e
police
costs th at result from th e need for increased
police presence
in areas of
greater gambling
activ ity.
One
significant problem
th at h as
plagued
th e
majority
of th e casino-crime liter-
ature is
analogous
to th e
problem present
in
calculating
th e
profit
and tax benefits of casi-
nos: T o estimate social
costs,
one sh ould not
count new crime around th e casino
only,
but
also consider w h eth er casinos reduced crime in
oth er locales
(for example,
th is could
h appen
if casinos mov e crime from oth er
locations).
Counting only
local crime as a cost is similar
to
counting only
local
profits
as a benefit.
T h e most
compreh ensiv e analysis
of th e
casino-crime link is Grinols et al.
(2000),
w h ich ev aluated
county-lev el
data for sev en
offenses in
ev ery
US
county
ov er 20
years,
and
controlled for about 50 v ariables. It concluded
th at on
av erage,
8-10% of crime in casino
counties in 1996 could be attributed to th e
presence
of casino
gambling
in th e
county,
resulting
in costs of $63
per
adult
annually
in
th ese counties. Furth ermore,
counties th at
border casinos also
experience
increased
crime rates,
w h ich
suggests
th at casinos
truly
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
150
Manage.
Decis. Econ. 22: 143-162
(2001)
increase
crime,
not
merely
sh ift it from one
location to anoth er. Estimates of th e cost of
non-Index crimes w ould add to total crime
costs. For
example,
insurance fraud is not an
FBI Index I crime. Estimates of th e fraud
committed
by gamblers
is $1.3 billion
per
year,29
or $6.61
per
adult
annually.3"
Studies of
problem
and
path ological gam-
blers h av e found similar effects.
Maryland
De-
partment
of Health and Mental
Hygiene (1990)
reported
th at 62% of
gamblers
in treatment
committed
illegal
acts as a result of th eir
gam-
bling,
80% committed civ il offenses and
23%'
w ere
ch arged
w ith criminal offenses. Lesieur
(1998b) surv eyed nearly
400 members of Gam-
blers
Anonymous,
57% of w h om admitted
stealing
to finance th eir
gambling.
On
av erage
th ese 400
people
stole
$135,000,
and th eir total
th eft w as ov er $30 million. Lesieur
(1992)
re-
ported
on
illegal
activ ities and civ il fraud en-
gaged
in
by path ological gamblers
to
gamble
or to
pay gambling
debts in fiv e
samples
from
h ospital inpatients,
Veterans Administration
and Gamblers
Anonymous groups,
male
pris-
oners,
female
prisoners,
and a female Gam-
blers
Anonymous sample
th at includes th e
w h ite-collar crime and oth er crimes listed in
item 1.3
2. Business and
Employment
Costs: T h ese costs
include lost
productiv ity
on th e
job,
lost time
and
unemployment:
sick
days
off for
gambling,
extended lunch
h ours,
leav ing early
to
gamble,
and
returning
late after
gambling.
Problem and
path ological gamblers
often
impose
costs on
th eir
employers
(in
addition to th eft or em-
bezzlement discussed in th e section on abused
dollars
below )
in th e form of an unreliable
presence
on th e
job
and reduced
productiv ity
w h en
present.
Betw een 21 and 36% of
problem
gamblers
in treatment
reported losing
a
job
because of th eir
gambling (Lesieur, 1998b).
Firing
an
employee imposes
costs on th e
w orker in terms of lost
output during
th e
period
of
unemployment
and on th e
employer
in terms of additional costs of
h iring
and
training
new
employees.
T h ese costs are
h igh er
th e
greater
th e
firm-specific
h uman
capital.
3.
Bankruptcy:
Law suits and
legal
costs, and bill
collection costs, bill collector h arassment are
among
th e
consequences
of
bankruptcy.
Path ological gamblers
often follow a
pre-
dictable
path
of
exh austing personal resources,
selling
insurance
policies, selling possessions,
and
'borrow ing'
from
family
and friends. T h eir
search for funds
may
lead th em to
acquire
multiple
credit cards th at
th ey
use to th e limit.
Debts w ill be
paid off,
of course,
w h en th e
indiv idual w ins
big
in h is next
gambling
episode. Bankruptcy
entails costs to creditors
attempting
to collect and costs to th e
legal
system
in court time and resources. SMR Re-
search
Corporation (1997, p. 118)
indicated
th at th ese costs
may
be
quite large,
'We set out
th is
year
to interv iew
many
of th e
leading
US
experts
on
gambling, gambling addiction,
and
th e financial
impacts
of
gambling.
T h eir stud-
ies h av e
suggested, fairly consistently,
th at
more th an 20% of
compulsiv e gamblers
h as
filed for
bankruptcy
as a result of th eir
gam-
bling
losses'.
4. Suicide: Lesieur
(1992)
concluded th at
problem
and
h av e
h igh er
suicide
rates th an th e
general public.32
Dozens of sto-
ries h av e been
reported
of
gamblers killing
th emselv es after
losing
at th e
casino,
some-
times on th e
premises.33
Consistent w ith
th is,
Ph illips
et al.
(1997)
found th at death s in Las
Vegas
w ere 2.5 times more
likely
to be a result
of suicide th an death s in oth er
comparably
sized
metropolitan
areas. Visitors to Atlantic
City
and Reno w ere 1.75 and 1.5 times more
likely
to die in suicides th an tourists to oth er
non-gambling
areas, and in Atlantic
City
th e
suicide rates did not become elev ated until
after casinos w ere introduced in 1978.
McCleary
et al.
(1998),
funded
by
th e Ameri-
can
Gaming
Association, contested
Ph illips'
findings.
Wh ile w e
recognize
th e
impact
of
casino
gambling
on suicide,
th e literature h as
not
prov ided sufficiently
reliable social cost
estimates, and, th erefore,
w e do not account
for such costs in th e table below .
5. Illness:
Among
th e forms of sickness associated
w ith
gambling
or affected
by
it are
depression,
stress-related illness, ch ronic or sev ere
h eadach es,
anxiety,
moodiness, irritability,
in-
testinal disorders, asth ma,
cognitiv e
distor-
tions,
and cardio-v ascular disorders.
Many
sickness costs are borne
by
th e
gambler,
but
th ey
also
appear
as resource costs w h en th e
gambler
seeks treatment. Gambler-borne costs,
ev en w h en not
absorbing
resources, h ow ev er,
are
tangible
costs to th e extent th at th e
gam-
bler w ould be
w illing
to
pay
to eliminate th e
problem.
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
151
Manage.
Decis. Econ. 22: 143-162
(2001)
6. Social Serv ice Costs: T h is
category
of costs
includes
th erapy/treatment
costs, unemploy-
ment and oth er social serv ice costs
(includes
w elfare and food
stamps).
7. Gov ernment Direct
Regulatory
Costs: Social
serv ice and
gov ernment
direct
regulatory
costs
are
paid primarily th rough
th e
gov ernment.
T h e
gambling industry
h as been
regulated
be-
cause it h as
h istorically
been
subject
to fraud
and abuse. Social serv ice costs transfer re-
sources from one
segment
of
society
to an-
oth er, consuming
resources in th e
process.
If
social costs include th e financial burden
placed
on th e
non-gambling society
th at w ould not be
present
in th e absence of
gambling,
th en th ese
costs sh ould be included for a
complete
assess-
ment of th e effects of
gambling. Regulatory
costs differ
by
state and
depend
on th e
type
of
casinos
(i.e. riv erboat,
Indian reserv ation,
etc.),
and extent of th e
responsibilities
of th e
regulatory agencies.
8.
Family
Costs: Families of
problem
and
path o-
logical gamblers
bear
gambling-related
costs
of
div orce, separation, spousal
abuse, and
ch ild
neglect. Alth ough
th ese costs are non-
pecuniary, th ey are, nev erth eless, tangible
and
real.
T h ey
can be
quantified
in terms of th e
amount of
money
an indiv idual w ould be w ill-
ing
to
pay
to remov e th e
problem.
In
practice,
such costs are
rarely
measured. Wh en social
serv ices become
necessary,
as w h en
gambling
leads to div orce
proceedings, th ey represent
resources lost to oth er uses in
society
and can
be measured
by
th e cost of th e serv ices
prov ided.
9. Abused Dollars: T h e final
category represents
lost
gambling money acquired
from
family,
friends,
or
employers
under false
pretenses.
T w o
examples
are
stealing
th at is nev er re-
ported
because th e th ief is a
relativ e,
and
money
'loaned' under duress th at is nev er
repaid.
Abused dollars
represent
costs to th e
non-gambling population.
T o th e extent th at
abused dollars
represent purch ases
of
gam-
bling
serv ices th at are
inefficiently sub-optimal
from th e
gambler's perspectiv e
or create mar-
ket
inefficiencies,
a
significant portion repre-
sents social costs to
society
as a w h ole ev en
allow ing
for
gains by
th e
gambler
or
gambling
sector.34
Social Cost Estimates T ied
Directly
to P&P
Gamblers
T able 2
reports
th e results of all
eigh t
studies th at
contain
original
research th at ties social costs
directly
to
path ological gamblers.35
T h e first tw o
row s sh ow th e location studied and th e
auth or(s),
respectiv ely.
T h e first column sh ow s th e
category
of
costs,
as outlined in th e
prev ious
section. T h e
studies are listed in order of date of
publication.
With th e
exception
of th e
path breaking paper by
Politzer et al.
(1981),
th e studies w ere
publish ed
betw een 1994 and 1999. T h e column totals
range
from a low of
$1,195 (Gerstein
et
al., 1999)
to a
h igh
of
$30,235 (Politzer
et
al., 1981).
T h e Execu-
tiv e Office of th e Gov ernor
(1994)
is th e
h igh est
post-1994
estimate. Because all studies omit some
of th e
costs,
th ese totals w ill understate th e actual
totals.
A
large
sh are of th e differences in th e totals is
explained by
differences in th e number of cost
components
th e studies estimated. T h e Executiv e
Office of th e Gov ernor
(1994)
estimated
only
crime costs in
Florida,
w h ile
T h ompson
and
Quinn (1999)
estimated ten
components.
T h e
study
w ith th e low est total cost (Gerstein
et al.,
1999)
estimated
only
four
categories.
No
study
estimated all th e
components.36 By far,
crime and
abused dollars are th e
largest
cost estimates. Ger-
stein et al.
(1999)
is th e
only study
th at
completely
omits crime
costs,
and
only
th e Executiv e Office
of th e Gov ernor
(1994)
and Gerstein et al.
(1999)
omit estimates of abused dollars. One
important
common ch aracteristic of all but one of th ese
studies is th at
th ey
are not
publish ed
in
peer-
rev iew ed
journals.
T h e Executiv e Office of th e
Gov ernor
(1994), Ryan
et al.
(1999), T h ompson
and
Quinn (1999)
and South Dakota
Legislativ e
Research Council
(1998)
w ere eith er
publish ed by
or
prepared
for state
agencies. T h ompson
et al.
(1996)
w as
publish ed by
th e Wisconsin
Policy
Research Institute. Politzer et al.
(1981)
w as
pre-
sented at th e Fifth National Conference on Gam-
bling
and Risk
T aking,
Gerstein et al.
(1999)
w as
presented
to th e NGISC. T h e
paper by T h ompson
et al.
(1998)
w as
presented
at th e T w elfth Na-
tional Conference on Problem
Gambling,
and
later
publish ed
in
Gaming
Research and Rev iew
Journal.
We used
many strategies
to ensure th at th e final
estimates of costs
per path ological gambler
w ere
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
152
Manage.
Decis. Econ. 22: 143-162
(2001)
T able 2. Annual Social Costs
per Path ological
Gambler
MD FL WI CT SD LA US SC Row
Politzer Exec.
T h ompson T h ompson
SD
Leg. Ryan
Gerstein
T h ompson av erages
et al. Office et al. et al. Research et al. et al. and
Quinn
for studies
(1981)
of Gov
(1996) (1998) Council
(1999) (1999) (1999)
1994-1999
($) (1994) ($) ($) 1998-1999 ($) ($) ($) ($)
($) ($)
Crime
Appreh ension
and increased
police
costs 44 71 1000 53 116 257
Adjudication (criminal
and civ il 1788 1234 994 27 649 476 676
justice costs)
Incarceration and
superv ision
costs 2828 15221 758 889 382 690 451 3065
Business and
employment
costs 11 265
Lost
productiv ity
on
job 1082 1082
Lost time and
unemployment 2717 3436 5936 320 2156 2913
Bankruptcy 515 118 316
Suicide
Illness
700 700
Social serv ice costs
T h erapy/treatment
costs 437 114 75 396 30 83 189
Unemployment
and oth er soc. sv c. 606 971 549 60 145 318 442
(incl.
w elfare and food
stamps)
Gov ernment direct
regulatory
costs
Family
costs
Div orce, separation 111 111
Abused dollars 14354 3802 9519 240 3175 2436 3834
13 586
CD
0
U,
0
0
0
w
C
tm
C,Q
_.
r3
0
r
-<
r
CA
0
n
r"
0-
rI
low er bounds.37 First, in
calculating
th e
av erage
annual cost
per path ological gambler by category
(sh ow n
in th e last column of T able 2 on th e
righ t)
w e omitted Politzer et al.
(1981).38
T h is
study
h ad
th e
h igh est
cost
estimates,
but w as conducted at a
different time and in a different
gambling
env iron-
ment from th e oth er studies.
Second, costs for
suicide and
gov ernment regulation
are
omitted,
because none of th ese studies estimated th em.
T h ird, w e did not
price adjust
th e
estimates,
but
rath er took th e v alues as
giv en by
th e auth ors.
Last,
many
studies combined th eir estimates for
path ological
and
problem gamblers.
We treated
th e numbers as if th e costs w e
report apply only
to
Because costs due to
are
h igh er
th an costs due to
problem gamblers,
th e estimates furth er under-
estimate th e costs connected to
path ological
gamblers.
T able 2 sh ow s th at th e total
av erage
social cost
of
eigh t
studies is $13,586
per path ological gam-
bler
per year.
If 1.5% of 196.65 million US adults
w ere
path ological gamblers,
th is w ould
imply
an-
nual social costs of $40.1 billion or $204
per
adult. If
are 1% of th e
population,
th e estimate reduces to $136
per
adult.
T able 3
replicates
T able 2 for
problem gam-
blers.
Only
Gerstein et al.
(1999)
and South
Dakota
Legislativ e
Research Council
(1998)
esti-
mated
any separate
costs
per problem gambler.
T h ese studies estimated
only
th ree of th e
many
cost
categories.
T h e
av erage
annual cost
per prob-
lem
gambler by
cost
category
is sh ow n in th e last
column. For th e same reasons discussed in
analyz-
ing
th e results for
th e
T able 3 total cost estimate of $912 due to
problem
gamblers
understates th e actual cost.
T able 4
applies
th e information in T ables 2 and
3 to
produce
annual national social costs
per
adult. T o test th e robustness of th ese cost esti-
mates,
w e use th e 95%, confidence bounds on th e
numbers of
problem
and
set
by
Sh affer et al.
(1997).39
T h is confidence
interv al sets th e fraction of
betw een 0.9 and 1.38%) of th e adult
popula-
tion, and th e fraction of
problem gamblers
betw een 1.95 and 3.65% of th e adult
popula-
tion. Based on th ese low er and
upper
bounds,
annual national social costs from
problem
and
T able 3. Annual Social Costs
per
Problem Gambler
US SD
Gerstein et
cl.
S. Dakota, Row
av erages:
(1999)
1998-1999 studies 1994-1999
($) ($) ($)
Crime
Appreh ension
and increased
police
costs
Adjudication (criminal and civ il
justice costs)
Incarceration and
superv ision
costs
Business and
employment
costs
Lost
productiv ity
on
job
Lost time and
unemployment
200 200
Bankruptcy
Suicide
Illness
Social serv ice costs
T h erapy/treatment
costs 360 360
Unemployment
and oth er soc. sv c. 155 549 352
(incl. w elfare and food
stamps)
Gov ernment direct
regulatory
costs
Family
costs
Div orce.
separation
Abused dollars
912
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
154
Manage.
Decis. Econ. 22: 143-162
(2001)
T able 4. National and
per
Adult Social Costs
NAT IONAL COST : BILLIONS of DOLLARS PER ADULT COST
Problem Rate
High
$30.6 $43.4 Problem Rate
High
$156 $221
Low l $27.5 $40.4
Low | $140 $205
Low
High
Low
High
Path ological
Rate
Path ological
Rate
Path ological
95% Confidence Bound: LOWER 0.9000%
Path ological
95% Confidence Bound: UPPER 1.3800%
Problem 95% Confidence Bound: LOWER 1.9500%
Problem 95% Confidence Bound: UPPER 3.6500%
path ological gambling range
from 27.5 billion to
ov er $43 billion. On a
per
adult basis, th e num-
bers
range
from a low of $140 to a
h igh
of $221.
Because Sh affer et al.
(1997)
estimated th ese con-
fidence bounds based on
samples
of th e nation
before th e time of
publication including
areas
w ith different
degrees
of casino
gambling th ey
clearly
understate th e fractions of th e entire US
population
th at w ould be identified as
path ologi-
cal or
problem gamblers
if casinos w ere
expanded
fully.
T h e costs of T able 4, th erefore,
also under-
state th e associated costs of full
gambling expan-
sion.
IMPLICAT IONS FOR FUT URE RESEARCH
T h is
paper
makes
many
contributions to th e dis-
cussion of social costs and benefits of casino
gaming,
and h as numerous
implications
for future
research in th is area.
First,
w e
prov ide
th e first
th eoretical
justification
of w h at sh ould be in-
cluded as costs and benefits. T h is
justification
is
based on indiv idual
utility
and
distinguish es
busi-
ness and social
profitability
for industries w ith
externalities. T h e lack of a clear th eoretical basis
h as
impaired
th e entire research
agenda
on th is
issue. Much research h as examined
relativ ely
mi-
nor issues or issues th at are not ev en
part
of a
properly
defined cost-benefit
analysis.
Con-
v ersely,
th ere are
relativ ely
few estimates of some
of th e
key components
of social costs and bene-
fits.
Consequently,
a
w ell-grounded
th eoretical
framew ork of costs and benefits w ill make future
research more
productiv e.
Second,
using
th is
th eoretically grounded
cost-
benefit
analysis
w e corrected sev eral common
conceptual
mistakes prev alent in th e casino and
gambling
literature. One
example
of a common
error is th e focus on local rath er th an total social
costs or benefits. On th e benefits
side,
increases in
local
profits
and taxes are often
w eigh ted h eav ily
w h ile losses in
profits
and taxes from
geograph i-
cally
distant areas are
w eigh ted
less or not at all.
Similarly,
on th e cost
side,
local crime is often
w eigh ted h eav ily
w h ile th ere is little discussion
about w h eth er crime w as
simply
mov ed from
oth er areas. Anoth er error is th e
frequent
use of
th e net
export-multiplier modeling
of
jobs,
an
inappropriate
meth od to determine social costs
and benefits.
Clearly, identifying
th ese errors w ill
reduce th em in future research .
T h ird,
w e used th e
th eory
to construct a clear
taxonomy
of benefits and costs as
applied
to th e
casino
industry.
T o estimate th ese costs and bene-
fits w e rev iew ed th e av ailable studies th at do
original
research on th is
topic.
T h is literature
sh ow s th at th e extreme
upper
bound on annual
total social benefits is $75
per
adult. T h e low er
bound for social costs,
based on th e estimates of
costs associated w ith
prev alence
of
problem
and
w as betw een $140-$221
per
adult.
Consequently,
th e av ailable research
indicates th at w h en
using
th e
h igh est
estimates of
benefits and th e low est estimates of costs, casino
gambling
fails a cost-benefit test
by
a ratio of 1.9
to one or
greater.40
Standard
Pigouv ian
correctiv e
th eory
for an
industry
w ith externalities is th at it
sh ould be taxed
by
an amount
equal
to th e costs
th at it
imposes
on
society.
Relativ e to th e
Copyrigh t
? 2001 Joh n
Wiley
& Sons, Ltd.
155
Manage.
Decis. Econ. 22: 143-162 (2001)
rev enues for a
representativ e
casino of about $230
per
adult each
year
from
nearby residents,
Pigou-
v ian correctiv e taxes w ould
represent
betw een 61
and 96% of casino rev enues.
Fourth ,
w e sh ow ed th at th e av ailable research
indicates th ere is a lack of
quality
research on both
th e benefit and cost sides of th e
debate, and th at
th ere is an
important
need for better research .
T h ere is a need for more
uniformity
in th e manner
in w h ich costs and benefits are treated. Peer-
rev iew -quality
studies not funded
by
th e casino
industry
or
by pro-
or
anti-gambling groups
are
especially
needed to refine and
improv e
th e cost-
benefit numbers th at are
currently
av ailable. T o
furth er refine th e cost-benefit
analysis
of casino
gaming
th e
follow ing questions
must be addressed.
Wh at is th e Effect of Casinos on th e Number and
Gambling
Patterns of Problem and
Path ological
Gamblers?
Because th e social costs of th e casino
industry
are
generated primarily by problem
and
path ological
gamblers,
it is essential to know h ow casinos affect
problem
and
T h ere is
abundant ev idence th at increased
gambling oppor-
tunities increase
problem
and
path ological gam-
bling.
For
example,
th e NGISC
reported
th at th e
presence
of a casino w ith in 50 miles
rough ly
doubled th e
prev alence
of
problem
and
path ologi-
cal
gambling.4'
Oth er indicators include th e
tremendous increase in th e numbers of
gamblers
seeking h elp
w h en casinos enter a
market,
th e
increase in
gamblers anonymous groups
w h en
gambling
enters a
state,
and th e ev idence from
surv ey
data on th e number of
problem
and
path o-
logical gamblers
before and after casino
expan-
sion.
Casinos
may
also affect th e amount of
gambling
by problem
and
An av er-
age
adult is
expected
to lose $200-300 each
year
in
casinos if
th ey
are
nearby,
w h ile a
typical path o-
logical gambler
often loses
10--20 times th is
amount. T h erefore, a small number of
path ologi-
cal
gamblers
accounts for a
significant portion
of
casino rev enues. A related issue is to determine th e
sh are of casino rev enues th at deriv e from
problem
and
Does th is sh are differ
by type
of
gambling?
For
example,
lotteries receiv e
a smaller
portion
of th eir rev enues from P&P
gamblers
because
lottery play
attracts a
larger
portion
of th e
population.
How much does an Additional Activ e Problem or
an Additional Activ e
Path ological
Gambler Cost
Society?
T h is
question
is best addressed
by studying prob-
lem and
path ological gamblers directly.
How ev xer,
estimates deriv ed from th is
sample may
be biased
because
only
a small fraction of P&P
gamblers
seek formal treatment. If th ose w h o seek
h elp
impose
th e
greatest
costs on
society,
our cost
estimates of P&P
gamblers
w ould be ov erstated.
Wh at is th e Life
Cycle
of a Problem and
Path ological
Gambler?
For
example,
w h en casino
gambling
becomes
av ailable for th e first
time,
w h at is th e beh av ioral
time
profile
for indiv iduals w h o enter and leav e
th e states of
problem
and
path ological gambling?
Do indiv iduals
begin
w ith a
period
of
increasing
gambling dependence,
mov e
th rough
a
period
of
problem gambling, progress
to
path ological gam-
bling,
seek treatment (or
w ith draw
unilaterally
from th e
problem).
and abstain th ereafter? Or are
th ere
relapses
and continued
problems
if treat-
ment is not
sough t.
T h is information could be
used to
predict
h ow
many currently
activ e
prob-
lem and
to
expect
for
giv en
population
as a function of th e
av ailability
of
casino
gambling.
Wh at Effect do Different
T ypes
of T reatment
h av e on Problem and
Path ological
Gamblers?
Such information w ould
h elp people
to know h ow
to
efficiently
allocate
funding
resources for treat-
ment interv entions.
How can Casino
Gambling
be Offered to
Minimize its Social Costs?
Quinn (2001)
discusses
many possible w ays
of
offering
casino
gaming
to reduce social costs. T o
ev aluate th e effectiv eness of th ese interv entions
and th eir
impact
on casino benefits one w ould
need to estimate th e
elasticity
of both P&P and
non-P&P
gamblers
to such actions.
Wh at are th e Net Profit and T ax Benefits of
Increasing
Casino
Gambling?
Rath er th an
estimating
a true social benefit, many
studies estimate
only
th e
gross
increases in
profits
or
only w eigh t
th e increased benefits to local firms
w h ile
ignoring
lost
profits
to oth er firms.
Copyrigh t
? 2001 Joh n
Wiley
&
Sons, Ltd.
156
Manage.
Decis. Econ. 22: 143-162
(2001)
Wh at are th e Distance Benefits of
Increasing
Casino
Gambling?
T o date
only
one
study
examines th is
important
question. T esting
th e robustness of th is result w ill
prov ide
more
insigh t
into th is understudied area.
Focusing
future research questions and
meth odologies
on a
clearly
formulated th eoretical
foundation w ill allow us to make our estimates of
both th e costs and benefits of casino
gaming
more
precise.
APPENDIX A
A
Study
of th e Economic
Impact
of th e
Gaming
Industry T h rough 2005, by
T h e Ev ans
Group:
A
Partial
Critique
International Game
T ech nology (+ IGT ),
a man-
ufacturer of
computerized
casino
gaming products
and v ideo
gaming mach ines,
and
operator
of
pro-
prietary gaming systems,
commissioned T h e
Ev ans
Group,
an econometric
consulting firm,
to
produce
a
study
of th e
impact
of th e
gambling
industry
in 1996. T h e 9
September
1996
press
release for th e
resulting report
entitled A
Study of
th e Economic
Impact of
th e
Gaming Industry
th rough
2005 issued
by
+
IGT
reported,
States and localities th at
permit
casino
gaming
h av e
improv ed
th eir ov erall economic
perfor-
Unemployment 9
Rate
8
7
6
5
4k
mance... T h e
study... reports
th at w h erev er
casino
gaming
h as been
implemented, employment
h as
risen, unemployment fallen,
and additional tax
rev enues h av e been
generated. (Emph asis added.)
T h e Ev ans
study
describes
impacts
for indiv idual
states. We w ill
briefly
examine th e
findings
related
to
Illinois,
a state w ith w h ich th e auth ors are
familiar. On
page
4-3 th e
report
states:
Based on th ese data,
it w ould
appear
th at th e
opening of
a casino reduced th e
unemployment
rate
in th at
county
in both th e
year
it w as
opened
and
in th e
follow ing year.
T h e
av erage employment
in
th ese
eigh t counties...implies
a total of 37 000 ex-
tra
jobs.
T h ese
multiplier figures
are much
h igh er
th an
ordinarily obtained,
and
employment
in th ese
counties
migh t
h av e risen for oth er reasons as
w ell. Noneth eless, th e
figures
do indicate th at
casino
gaming
h as been a boon to th ese
counties,
especially
th ose th at are more rural.
(Emph asis
added.)
Most casinos
opened
after 1991. T h e
period
1991-1996 cov ered
by
th e
study, th erefore,
coin-
cided w ith th e nationw ide economic
expansion
coming
out of th e recession of 1990-1991. Em-
ployment
w as
rising
and
unemployment
w as
falling
in
many counties,
w ith or w ith out th e
introduction of casinos. T h e
auth ors, th erefore,
w ere
righ t
to feel
uneasy.
T h eir caution th at 'em-
ployment
in th ese counties
migh t
h av e risen for
oth er reasons' sh ow s
th ey
knew th at
simple
be-
fore-and-after
comparisons finding declining
un-
employment
and
increasing employment prov ed
LlIi i
HILL
4s
%
/s,
4,
v
% 4V4
Unemployment
rate 1 at year of casinos
M
1994
unemployment
rate
Figure
1. T h e Ev ans
Group study, reproduced figure
4-1.
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
157
Manage.
Decis. Econ. 22: 143-162
(2001)
'%t %,?
,<%,
dN
"
Figure
2. Casino counties are
indistinguish able
from non-casino counties.
noth ing
about th e effects of casinos in a
country
recov ering
from recession.
Figure
1
reproduces
figure 4-1, prov ided
in th e
original study.
T h e
auth ors
explain
th at th e observ ed
drop
in casino
county unemployment
rates exceeded th e state
av erage by
0.3 and 0.2
percentage points
on av er-
age
in th e first and second
year
after introduction.
T h e auth ors' conclusions are noted abov e. T h e
rest of th e
story
is
prov ided
below .
T h e
study giv es
th e
impression
th at counties
th at
opened
casinos
experienced
better economic
performance
th an th ose th at did not.
How ev er,
Illinois contains 102 counties. We can select oth er
counties th at h ad th e same
unemployment
rate
(w ith in
0.1
percentage point)
as th e casino
county
in th e inital
period
and
compare
th eir
peformance
directly.
T h is is done in
Figure
2. As sh ow n
th ere,
th e
unemployment
rate
dropped
in all counties
w ith similar initial
unemployment.
Some counties
did better th an casino
counties,
some counties did
w orse. From left to
righ t,
bottom row
first,
th e
casino counties are numbers
6, 1, 3, 2, 2, 3, 7,
3.
Nineteen counties
performed
better th an th eir
casino
cousin,
w h ile 19
performed
w orse.
A statistical test confirms th at th e
drop
in
unemployment
of casino counties is
statistically
insignificantly
different from th e
drop experienced
by
th e
comparable
non-casino counties sh ow n in
Figure
2. Let AU denote th e
ch ange
in
county
unemployment
rate minus th e
ch ange
in state
unemployment
rate for th e same
period,
and let
Casino
identify
counties th at introduced casinos in
th e initial
period (Casino
=
1 if a
county
intro-
duced a casino, 0
oth erw ise).
T h en
running
th e
follow ing regression,
AU =
a + b Casino + e
rev eals th at coefficient b is 2.75
(consistent
w ith
th e 0.2 and 0.3
percentage point
differences re-
ported by
T h e Ev ans
Group),
but w ith a standard
error of 0.856
implying
a
p-v alue
of 0.4. Coeffi-
cient b
is, th erefore, statistically indistinguish able
from 0 at conv entional lev els.
NOT ES
1.
Gambling
rev enue is th e net amount of
money
th at
th e
gambling operator
extracts from
patrons.
It
equals
th e 'h andle'
(gross
amount
w agered-
w h ich
may
reflect th e same
ch ips being
bet
many
times before it is
ultimately
retained or
lost)
less
payouts, prizes,
or
w innings
returned to
players.
For
example,
if
players w ager
$1 000 000 on out-
comes of a roulette w h eel ov er th e course of an
ev ening,
and $880 000 is returned to th em as w in-
nings (some
roulette slots are reserv ed for th e
h ouse),
th en
operator
rev enue is $120 000.
2.
According
to th e Indian
Gaming Regulatory
Act of
1988,
Class I
gambling
consists of 'social
games
solely
for
prizes
of minimal v alue'. Included in
Class I
gambling
are traditional Indian
games
iden-
tified w ith tribal ceremonies and celebrations. Class
II
gambling
includes
bingo
and
'games
similar to
bingo'.
Class III
gambling
includes 'all forms of
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
158
Manage.
Decis. Econ. 22: 143-162
(2001)
gaming
th at are not Class I
gaming
or Class II
gaming',
such as
blackjack,
slot
mach ines, roulette,
and oth er
casino-style games.
3. Kindt
(1994),
Grinols
(1996),
Grinols and Omorov
(1996)
and Henriksson
(1996)
discussed a number
of th ese.
4. Public Law 104-169 of th e 104th
Congress
estab-
lish ed th e NGISC. For more information about its
mission, composition
and
findings
see
h ttp://
w w w .ngisc.gov /.
5. We sh ow below th at both
concepts
are
necessary
to
a
proper
cost-benefit assessment of casinos.
6. We w ill sh ow below th at it sh ould be
present.
7. We follow
th rough out
th e
paper
standard
general
equilibrium accounting
conv entions for
describing
inputs
and
outputs
in
consumption
and
production.
8. T h at
is, $100 of
utility
is defined to be th e
utility
th at can be ach iev ed
by optimally spending
$100 at
prices p,
w ith nearest casino d miles
aw ay.
9. For
example,
th e returns to
playing
roulette, slot
mach ines,
or a
blackjack game
are
approximately
th e same
regardless
w h ere offered. T h e framew ork
could be modified to allow for different
qualities
of
gambling.
In th is case th e model w ould deal w ith
multiple, imperfectly
substitutable
goods.
10. T h e transfer of w ealth in
gambling
is
generally
from
relativ ely poor
to
relativ ely w ealth y.
T h ere-
fore,
if a dollar
generates
more
utility
for rich th an
poor,
our
assumption
understates th e social bene-
fits. If a dollar
generates
more
utility
for th e
poor
th an th e
rich ,
our
assumption
understates th e social
costs of casinos.
11. A
positiv e
element of
yj
denotes
output
of a
good
or
serv ice,
and a
negativ e component
denotes th e
use of an
input.
12.
Alth ough
it is not central to our
objectiv e
in th is
paper,
w e include z to be consistent w ith th e
gen-
eral framew ork w e
dev elop. Excluding
z does not
affect th e central
arguments
of th is
paper. Compo-
nents of z are
economy
excess demands for traded
goods.
A zero denotes a non-traded
good,
w h ile a
positiv e entry
denotes
imports.
13. Use th e fact th at
1i ,i
= 1.
14. In
1998, gambling
rev enues w ere
approximately
0.5% of GDP and casino rev enues w ere
approxi-
mately
0.25%.
15. It is
conceiv able,
of
course,
in certain circumstances
th at th e introduction of casinos could
ch ange prices
enough
to matter to local residents. For
example,
if
casinos increased
employment
and th e local
popu-
lation,
th e demand for local
h ousing
w ould in-
crease,
th us
raising h ousing
prices and
creating
capital gains
for residents. In such
cases, h ow ev er,
th e reduction in demand for residential
property
and
capital
losses in th e areas from w h ich th e new
residents came w ould h av e to be taken into ac-
count. Ov er time,
if new
h ousing responded
to th e
increased demand, th e
prices
of th e
existing
stock
of
h ousing
w ould decrease. Because
gambling
doesn't create new
people,
but
only
mov es th em
from one
place
to anoth er,
a reasonable first
ap-
proximation
is th at th e net effect of
gambling
on
capital gains
and consumer
surplus
considerations
w ould be small.
16. We
presume
th at th e
jobs being compared
in tw o
locations are
comparable.
Blair et al.
(1998) argued
th at
'employees
in
gaming industry occupations
are
less satisfied w ith th eir
jobs
th an th ose in oth er
industries'. If
jobs
are
dif/frent
in tw o locations,
th en th e
jobs
w ould
appear
in th e formula as
different because w orkers w ould demand
compen-
sating w age differentials,
and th is w ould affect
profitability.
If
compensating w age
differentials do
not arise, but w orkers face non-market constraints
th at cause th em to w ork h ours th at are not
optimal
giv en
th e
w ages paid,
th ese costs w ould
appear
in
th e
unemployment
terms of
Equation (5).
17. For
industry
rev enue
data,
see International
Gaiming
and
Wagering
Business
(1999, p. 24).
18. T h e v alue of tobacco
grow n
each
year
is $39 bil-
lion. Encarta
Encyclopedia, h ttp://encarta.msn.com/
find/Concise.asp?ti
= 02A43000 # s12.
19. T h e
proposal
w as for West Dundee, Illinois. T h e
study reported,
'Both
Ch ristiansen/Cummings
and
Mirage
Resorts estimate local
gaming
demand
by
applying gaming
w in
per capita
factors to th e
popu-
lation
residing
w ith in concentric circles of a
gaming
v enue. T h e factors decline as distance increases.
T h e $200 w in
per capita applicable
to th e 0-50 mile
segment
w as
dev eloped jointly by representativ es
of
Mirage
Resorts and Dr
Cummings
to
apply
to th e
local
population
in th e New Orleans env irons in a
1992 ev aluation of th e New Orleans
gaming
market.'
20.
See,
for
example,
GLS Research
(1994)
Clark
County (Las Vegas, Nev ada)
Residents
Study
1993-1994. Ev en in Las
Vegas,
one-th ird of th e
population
does not
gamble.
21. For
example,
a
study
of
w agers
in Minnesota
(Smith
and
Craig, 1992; T ice, 1995)
found th at 1%
of
gamblers
accounted for 50%) of
w agers,
and th at
10% accounted for 80%. An Illinois
study (Gazel
R,
T h ompson
WN. 1996. Casino
gamblers
in Illi-
nois: w h o are
th ey? Manuscript,
1-25
(plus
data
supplied by
th e
auth ors))
found th at 10% of bettors
accounted for 66% of
w agers. Heav y gambling
is
not th e same as
problem
and
path ological gambling
ev en
th ough
th e rev enues of P&P
gamblers figure
disproportionately among
th e rev enues of th e
h igh est-gambling segment
of th e
population.
Wh en
compared
w ith th e
population
at
large,
th e amount
gambled by
P&P
gamblers implies
th at th e sh are of
casino rev enues from
problem
and
path ological
gamblers
can be as much as 1/4
to
1/2
of casino
rev enues (see Grinols and Omorov , 1996).
Lesieur
(1998b) reported
th at 48.7% of casino rev enues in
Nov a Scotia came from
problem gamblers,
and
th at 55% of rev enues for casino cards and dice
games
came from
problem gamblers
in
Wash ing-
ton. In oth er locations h e found th at
percentages
ranged
betw een 26.7 and 41.4%. In
Montana, 37%
of th e rev enues of v ideo
gambling
mach ines w as
estimated to come from
problem
and
path ological
gamblers (Polzin
et al., 1998).
T h e
Productiv ity
Copyrigh t
? 2001 Joh n
Wiley
& Sons,
Ltd.
159
Manage.
Decis. Econ. 22: 143-162
(2001)
Commission
(1999) reported
th at
problem gamblers
account for 2.1% of th e adult
population
but one-
th ird of all
gambling
rev enues in Australia.
Volberg
et al.
(2001)
also examined th e distribution of rev -
enue from different
types
of
gamblers.
22. T h is
description applies
in Illinois and
many
oth er
midw estern states. In
Minnesota,
for
example, only
American Indians
operate
casinos. In locations
such as Atlantic
City
or th e Gulf Coast of Missis-
sippi, regulations
allow
entry
to all as
long
as
certain
operating requirements
are met. In th ese
locations
competitition
driv es economic
profits
to
zero.
23. Non-Indian casinos
paid
ov er $2 billion in taxes to
th e v arious states on
gaming
rev enues in 1997. CT 's
tw o Indian casinos
paid
$236 million to th e state
th at
year.
In
comparison,
states
generated
rev enues
of
approximately
$10 billion from net
proceeds
of
lotteries in
1997,
or $51.15
per
adult.
24. See Economic
Report
of th e
President, 1999,
T able
B-15,
column 8.
25. How sh ould w e treat demand deriv ed from addic-
tion? If addiction is not rational th en its deriv ed
demand sh ould be treated
differently. We,
th ere-
fore, report
both
figures
abov e. In th e low er
figure,
w e assumed th at 32% of casino rev enues are from
P&P
gambling.
26. A
surv ey
of th is literature and list of references can
be found in Adam Rose and Associates
(1998)
and
th e NGISC
(1999), appendix
5 on Economic
Dev elopment.
27. Lev en et al.
(1998) prov ide
an
example
of h ow th e
focus on
job
creation
may
mislead th e
unw ary
or
untrained.
T h ey w rote,
'T h is
study
seeks to take an
objectiv e
look at th e
economic
impact
of th e
gaming industry
on th e
Missouri
economy.
Wh ere do th e
gaming
rev -
enues come from? How are
th ey
redistributed in
th e
economy? By
h ow much do state and local
gov ernments
benefit? Wh at is th e net bottom-line
economic
impact?... [Gaming]
does add
spend-
ing, income,
and
jobs
to th e Missouri
economy.
It sh ould be addressed in th is context.'
Wh ile th e auth ors do not claim th at th e answ ers to
th eir
questions
constitute a cost-benefit
ev aluation,
th eir
plea
th at
gambling
adds
'spending, income,
and
jobs
to th e Missouri
economy'
and th at 'it
sh ould be addressed in th is context' could
easily
be
misinterpreted
to mean th at a calculation of in-
come, jobs,
and
employment
is
synonymous
w ith a
cost-benefit ev aluation. In th eir
summary (p. 75)
th ey
w rote:
'T h e focus of th is
study
h as been th e determina-
tion of w h eth er net new
output
(and
jobs
and
employment)
h av e been created state-w ide in
Missouri as a
consequence
of casino
gaming
operations,
and if so h ow much ... T h e 'bottom
line' is th at
significant
additions to th e Missouri
economy
h av e been ach iev ed. As of
1997,
almost
18 000 net new
jobs,
$500 million in added
per-
sonal income, and ov er $750 million of added
output
h av e benefited th e state's
economy.'
Wh o w ould
argue
w ith such
figures?
Or be aw are
th at
regardless
of th eir
accuracy,
casinos in Mis-
souri
migh t
fail to
pass
a cost-benefit test and th us
be h armful to state w elfare?
28. See Grinols et al.
(2000)
for a
complete
rev iew of
th is literature. Each of th e
follow ing
crimes h as
been
alleged
in th e literature to be associated w ith
gambling.
Index I Violent Crime
(Aggrav ated
As-
sault, Robbery, Rape, Murder), Property
Crime
(Larceny, Burglary,
Auto
T h eft),
and non-Index I
crime such as Embezzlement and
Employee T h eft,
Loan
Fraud,
Insurance
Fraud, Forgery (including
ch eck
forgery),
T ax
Ev asion,
T ax
Fraud,
Con
Games (Sw indles, Hustling Cards,
Dice or Oth er
Games), Bookmaking, Working
in an
Illegal Game,
Pimping,
Prostitution, Selling Drugs,
and
Fencing
Stolen Goods.
29. Lesieur
(1992, p. 45)
and Executiv e Office of th e
Gov ernor
(1994, p. 67).
30. National
population
data
by age
coh ort are on th e
US Census Bureau w ebsite
h ttp://w w w .census.gov /
population/estimates/nation/intfile2-l.txt.
As of 25
August 2000,
th e US h ad a
population
of
275 130 000.
Moreov er,
196 649 000 w ere
aged
20
or older.
31. See T able 2.
32. See also Frank et al.
(1991).
33.
Representativ e
of such cases is th e
follow ing
ac-
count,
'A Florida man w h o lost about $50000
w h ile
gambling
h ere
(Atlantic City) during
th e
past
tw o
days
died
T uesday
after h e
jumped
sev en floors
from a
T rump
Plaza Hotel and Casino roof onto
Columbia
Place,
officials said'. Brian
Hickey,
Staff
Writer,
18
August 1999,
South
Jersey Publish ing
Co.
34. T h e minimum social costs of th is
category
are th e
v alue of th e resources
spent by
th ose
trying
to steal
and cov er
up
th eir offenses and th e v alue of th e
resources
spent by potential
v ictims to decrease
th eir likelih ood of
being
v ictimized. T h ere
may
be
anoth er
component
of cost in
addition,
h ow ev er.
Social costs can be
h igh er
if th e
original
ow ners of
th e
property
v alue it more th an th e offenders do.
For
example,
if th e ow ners v alued th eir
property
at
$1000 and th e offenders w h o stole it sold it to
someone w h o v alued it at
$300,
th ere w ould be an
additional social loss of $700. Furth ermore,
if th e
th ief is a
path ological gambler
and
spends
th e
w rongly acquired
$300
gambling,
h is
expenditures
may
reflect addiction rath er th an rational ch oice.
In th at case th ere w ould be social cost
equal
to
some or all of th e $300 because of h is
sub-optimal
allocation of resources to th e
gambling
sector. Last,
alth ough
th ere is some debate about w h eth er to
count stolen dollars as costs to all of
society (w h ich
includes th e
th ief)
because 'th e th ief
gets
th e
money',
it is clear th at th e
non-gambling portion
of
society
w ill be made w orse off
by
such
actions,
and
losses to th e rest of
society
are
important
in th e
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
160
Manage.
Decis. Econ. 22: 143-162
(2001)
policy
debate because
th ey suggest
th at all of th e
abused dollars
represent
social costs to th e non-
gambling
sector.
35.
Westph al
et al.
(1999)
is not used in T able
2,
but
supplements Ryan
et al.
(1999).
T h e South Dakota
Research Council
study
w as
completed
in
1998,
but
addenda w ere added in 1999. See also Finance and
Administrativ e Cabinet,
Commonw ealth of Ken-
tucky (1999),
Florida
Department
of Law Enforce-
ment
(1994),
Florida Sh eriffs Association
(1994),
Iow a
Racing
and
Gaming
Commission
(1995),
Lesieur
(1998a).
36. As an alternativ e
w ay
of
sh ow ing
th at th e differences
in th e totals are driv en
largely by
th e number of cost
categories estimated,
w e
compared
th e totals after
'filling
th e
gaps'
in each
study using
th e
av erage
cost
for a
giv en category
from th ose studies th at did
estimate th ose
particular
costs. Wh en
doing so,
th e
v ariance in th e totals decreased
substantially.
T h e
low est totals w ere for South Dakota
Legislativ e
Research Council
(1998-1999)
and
T h ompson
and
Quinn (1999),
$7396 and
$8047, respectiv ely.
T h e
largest
w ere $25 742
by
th e Executiv e Office of th e
Gov ernor
(1994)
and $18 203
by T h ompson
et al.
(1998).
37. In addition to our use of th e
numbers,
some
studies,
such as
T h ompson
et al.
(1998) intentionally
formed
th eir
original
estimates
conserv ativ ely
to understate
costs.
38.
Including
th e nominal v alue of th is
study
w ould
increase th e cost estimate for th ree of th e four costs
it estimates.
Using
th e v alues
adjusted
for 19
years
of
price
lev el
ch anges
w ould h av e
significantly
in-
creased th e estimates of all four costs.
39. See T able
5, p.
34.
40. Our
h igh est
estimate of benefits w as
$75;
our low est
estimate of costs $140.
Applying
th e
per
adult costs
of $221 from T able 4 to th e estimate of benefits
adjusted
for P&P
gamblers
of $34
implies
th at casino
gambling
fails a cost-benefit test
by
a ratio of 6.5:1.
41. NGISC
(1999, p. 4-4).
REiERENCES
Adam Rose and Associates. 1998. T h e
regional
economic
impacts of
casino
gambling:
Assessment
of
th e literature
and establish ment
of
a reasearch
agenda. Report pre-
pared
for th e National
Com-
mission
(August).
Blair
BF,
Sch w er
RK, Waddoups
JC. 1998.
Gambling
as an economic
dev elopment strategy:
th e
neglected
issue of
job
satisfaction and
nonpecuniary
income. T h e
Rev iew
of Regional
Studies
28(1):
47-62.
Deloitte and T ouch e. 1992. Economic and oth er
impacts
of a
proposed gaming,
entertainment and h otel
facility.
Prepared
for th e
City
of
Ch icago Gaming
Commis-
sion.
Eadington
WR. 1999. T h e economics of casino
gam-
bling.
Journal
of
Economic
Perspectiv es 13(3):
173-
192.
T h e Ev ans
Group.
1996. A
study
of th e economic
impact
of th e
gambling industry th rough
2005.
Study
commis-
sioned
by
International Game
T ech nology, Septem-
ber,
iii + 105.
Executiv e Office of th e Gov ernor. 1994. Casinos in
Florida. Office of
Planning
and
Budgeting:
T h e
Capi-
tol, T allah assee,
FL.
Finance and Administrativ e
Cabinet,
Commonw ealth of
Kentucky.
1999. Assessment of th e economic and
social
impact
of
expanded gaming
in th e common-
w ealth and
neigh boring region. December,
iii + 182.
Florida
Department
of Law Enforcement. 1994. T h e
Question
of
Casinos in Florida: Increased Crime: Is it
Worth th e Gamble? State of Florida:
T allah assee,
FL.
Florida Sh eriffs Association. 1994. Casinos and crime: is
it w orth th e
gamble?
A
summary report
and
position
paper.
Florida Sh eriffs Association:
T allah assee,
FL.
Frank
ML,
Lester
D,
Wexler A. 1991. Suicidal beh av ior
among
members of
gamblers anonymous.
Journal
of
Gambling
Studies 7: 249-254.
Gerstein
D, Murph y S,
T oce
M, Volberg R,
Harw ood
H,
T ucker
A,
Ch ristiansen
E, Cummings W,
Sinclair
S. 1999.
Gambling impact
and beh av ior
study: report
to th e National
Commission.
National
Gambling Impact Study Commission, (1
April).
GLS Research . 1994. 1993-1994 Clark
County
Residents
Study.
Las
Vegas
Conv ention and Visitor's
Auth ority:
Las
Vegas,
NV.
Grinols EL. 1996. Incentiv es
explain gambling's grow th .
Forum
for Applied
Research and Public
Policy 11(2):
119-124.
Grinols EL. 1999. Distance effects in
consumption.
Rev iew
of Regional
Studies
29(1):
63-76.
Grinols
EL,
Mustard
DB, Dilley
CH. 2000.
Casinos,
Crime,
and
Community
Costs.
Univ ersity
of Illinois:
Univ ersity
of
Georgia;
1-35.
Grinols
EL,
Omorov J. 1996.
Dev elopment
or dreamfield
delusions?:
assessing
casino
gambling's
costs and ben-
efits. T h e Journal
of
Law and Commerce
16(1):
49-87.
Henriksson LE. 1996.
Hardly
a
quick
fix: casino
gam-
bling
in Canada. Canadian Public
Policy XXII(2):
116-128.
International
Gaming
and
Wagering
Business. 1999.
August.
Iow a
Racing
and
Gaming
Commission. 1995. Des
Moines
Register.
13 October.
Ch ristiansen/Cummings
Associates:
Arlington,
MA.
Kindt JW. 1994. Increased crime and
legalized gambling
operations:
th e
impact
on th e socio-economics of
business and
gov ernment.
Criminal Law Bulletin. 43:
538-539.
Lesieur H. 1992.
Compulsiv e gambling. Society 29(4):
43-50.
Lesieur H. 1998a.
T estimony
before th e National Gam-
bling Impact Study Commission,
Atlantic
City,
New
Jersey,
22
January 1998, Institute of Problem
Gambling.
Lesieur H. 1998b. Costs and treatment of
path ological
gambling.
T h e Annals
of
th e American
Academy of
Political and Social Science
(Gambling:
socioeco-
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
161
Manage.
Decis. Econ. 22: 143-162
(2001)
nomic
impacts
and
public policy, Frey JH, special
editor)
556: 153-171.
Lev en
C,
Ph ares
D,
Louish omme C. 1998. T h e eco-
nomic
impact
of
gaming
in Missouri.
Report
to Civ ic
Progress,
St.
Louis, April.
Maryland Department
Hygiene,
Alcoh ol and
Drug
Abuse Administration. 1990. T ask
Force on
Gambling
Addiction in
Maryland. Maryland
Department
Hygiene,
Alcoh ol
and
Drug
Abuse Administration:
Baltimore,
MD.
McCleary R,
Kenneth
C, Feng W,
Merrill
V, Napoli-
tano
C,
Moles
M,
Graffeo B. 1998. Suicide and
gambling:
An
analysis
of suicide rates in US counties
and
metropolitan
areas.
Report
to th e American
Gambling
Association.
Manuscripts.
Mirage
Hotel. 1993.
Response
of th e
dev elopers
of th e
proposed
Fox Riv er resort to th e West Dundee riv er-
boat
project
task force.
National
Commission. 1999.
National
Commission Final
Report w w w .ngisc.gov [accessed
18 June
1999].
Ph illips DP, Welty WR,
Smith MM. 1997. Elev ated
suicide lev els associated w ith
legalized gambling.
Sui-
cide and
Life-T h reatening
Beh av ior 27: 373-378.
Politzer
RM,
Morrow
JS, Leav ey
SB. 1981.
Report
on
th e societal cost of
path ological gambling
and th e
cost-benefit/effectiv eness
of treatment. T h e Joh ns
Hopkins Compulsiv e Gambling Counseling Center,
Presented at Fifth National Conference on
Gambling
and
Risk-T aking.
Polzin
PE, Baldridge J, Doyle D, Sylv ester JT , Volberg
R,
Moore WL. 1998. Final
Report.
Presented to th e
Montana
Gambling Study Commission,
30
September
1998,
Bureau of Business And Economic Research
(Univ ersity
of
Montana-Missoula).
Productiv ity
Commission. 1999. Australia's
gambling
industries. Final
Report, Summary (26 Nov ember,
Report
No.
10).
Quinn
FL. 2001. First do no h arm: Wh at could be done
by
casinos to limit
path ological gambling. Manage-
rial and Decision Economics 22: 133-142.
Ryan T P, Speyrer JF,
w ith Beal
ST ,
Burckel
DV,
Cunningh am
BR, Kurth MM, Scott
LC,
Wall
JL,
Westph al
JR. 1999.
Gambling
in Louisiana: a
benefit/
cost
analysis. Prepared
for th e Louisiana
Gaming
Control Board,
Louisiana State
Univ ersity
Medical
Center
(April).
SMR Research
Corporation.
1997. T h e Personal
Bankruptcy
Crisis, 1997:
Demograph ics, Causes,
Im-
plications
& Solutions. SMR:
Hackettstow n, NJ.
Sh affer
HJ,
Hall
MN,
Vander Bilt J. 1997.
Estimating
th e Prev alence
of
Disordered
Gambling
Beh av ior in th e
United States and Canada: A
Meta-Analysis.
Harv ard
Medical Sch ool: Boston.
Smith
F, Craig
WJ. 1992. Wh os in for h ow much ?
CURA
Reporter.
Center
for
Urban and
Regional
Affairs
22:
11-14,
16
(plus
data
supplied
to th e
auth ors).
South Dakota
Legislativ e
Research Council. 1998. Eco-
nomic and
fiscal impacts of
th e South Dakota
gaming
industry.
1-146+ Attach ment: Letter of 7
January
1999 to Mr.
T erry Anderson, Director,
South Da-
kota
Legislativ e
Research
Council, T able South Da-
kota T otal Estimated Incremental Social Costs.
T h ompson WN,
Gazel
R,
Rickman D. 1996. T h e social
costs of
gambling
in Wisconsin. Wisconsin
Policy
Research Institute
Report 9(6):
1-44.
T h ompson WN,
Gazel
R,
Rickman D. 1998. Social
costs of
gambling:
A
comparativ e study
of
nutmeg
and ch eese state
gamblers.
T w elfth National Confer-
ence on Problem
Gambling,
NV 1998.
Gaming
Re-
search and Rev iew Journal
5(1):
1-15.
T h ompson WN, Quinn
FL. 1999. An economic
analy-
sis of mach ine
gambling
in South Carolina. T h e
Education Foundation of th e South Carolina
Policy
Council, May
1999.
T ice DL. 1995.
Wh y
th e
press
h ates
gaming.
Casino
Executiv e
August:
57-58.
Volberg R,
Gerstein
DR,
Ch ristiansen EM.
Baldridge
J. 2001.
Assessing self-reported expenditures
on
gam-
bling. Managerial
and Decision Economics 22: 77-
96.
Westph al JR,
Joh nson
LJ,
Stev ens L. 1999.
Estimating
th e social costs of
gambling
disorders in Louisiana
for 1998.
Study
Performed Under Subcontract From
Univ ersity
Of New
Orleans,
UNO # 313-204-124
(23
March ).
Copyrigh t
? 2001 Joh n
Wiley
&
Sons,
Ltd.
162
Manage.
Decis. Econ. 22: 143-162
(2001)

Cazinouri Canada

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Cazinouri Canada

Загружено:

Авторское право:

Доступные форматы

Business Profitability versus Social Profitability: Evaluating Industries with Externalities, the

Вам также может понравиться