FAMILY AND Sharpe
Abdel-Ghany,
CONSUMER
/ LOTTERY
SCIENCES
EXPENDITURES
RESEARCHIN
JOURNAL
CANADA
Lottery Expenditures in Canada: Regional
Analysis of Probability of Purchase,
Amount of Purchase, and Incidence
Mohamed Abdel-Ghany
University of Alabama
Deanna L. Sharpe
University of Missouri
This article has two purposes: First, to examine the effect of household characteristics on lottery
expenditures in six regions of Canada using a double hurdle model to distinguish between the
decision to play and the decision of how much to spend. Second, to estimate the incidence of lottery expenditures. Using the 1996 Canadian Family Expenditure Survey, the results portray the
profile of households that have the probability of becoming participants in lottery play as well as
the profile of households that spend more on lottery purchases. Lottery expenditures are found to
be regressive in all regions.
In the past three decades since lotteries were introduced in Canada,
they have grown into a multibillion-dollar industry. During fiscal
year 2000, Canadian lottery sales reached $9.1 billion (Canadian dollars). Per capita sales was $291 in Canada. Quebec led the provinces
with fiscal sales of $3.4 billion, followed by Ontario ($2.2 billion). On a
per capita basis, the Canadian leader in lottery sales was the
Lotto-Quebec lottery ($465) (North American Association of State
and Provincial Lotteries [NASPL], 2000a).
In Canada, lotteries are overseen by provincial governments and
are marketed and distributed collaboratively by a cartel of provincial
agencies. In the Maritimes, for example, lotteries are managed and
conducted under the auspices of the Atlantic Canada Lottery Corporation that was established in 1976, comprising representation from
each of the four Maritime Provinces. The Western Canada Lottery
Corporation operates in a similar fashion by coordinating the efforts
of Alberta, Manitoba, Saskatchewan, Northwest Territories,
Nunavut, and Yukon Territory and was established in 1974. Ontario,
Quebec, and British Columbia each operate their own lottery entities,
Family and Consumer Sciences Research Journal, Vol. 30, No. 1, September 2001 64-78
© 2001 American Association of Family and Consumer Sciences
64
Abdel-Ghany, Sharpe / LOTTERY EXPENDITURES IN CANADA
65
which were established in 1975, 1969, and 1985, respectively. All provinces combine at the national level for Canada-wide lotteries through
the jointly owned Inter-Provincial Lottery Corporation (NASPL,
2000b).
Canadian jurisdictions promote gambling as an economic development tool that creates jobs, funds charitable groups, and augments
provincial coffers. Lottery proceeds benefit different programs in different jurisdictions. In many cases, lottery profits are combined with
tax and other revenues in a government’s general fund. In other cases,
lottery proceeds are set aside for a specific cause, such as education,
cultural activities, economic development, environmental programs,
health care, sports facilities, tax relief, programs for the elderly, and
others.
One image of the legal lottery industry sometimes advanced by its
critics is that it preys on ignorance or, at best, the false hopes of those
who live at the lower end of the Canadian economic spectrum.
Despite the deep-seated moral and ethical arguments that may be put
forward by local groups and individuals, it is most helpful to move
the debate to the level of factual information. Most important is the
extent to which various socioeconomic and demographic variables
are instrumental in determining the probability of lottery purchase as
well as the amount of purchase. One public policy issue stemming
from the lottery is tax regressivity (lottery expenditures as a percentage of income fall as income increases), and related to this is the fear
that some households, especially the poor, overspend on lottery purchases and become impoverished due to states’ promotion of
gambling.
The purpose of this study is twofold: (a) to examine the effects of a
number of socioeconomic and demographic variables on the probability of lottery purchase as well as the amount of purchase in the six
regions of Canada, and (b) to determine the extent to which lottery
purchases are regressive in their incidence on taxpayers.
REVIEW OF LITERATURE
Most previous studies have focused on determining the economic
burden of the implicit lottery tax (Borg & Mason, 1988; Brinner &
Clotfelter, 1975; Clotfelter, 1979; Clotfelter & Cook, 1987, 1989;
Hansen, 1995; Hansen, Miyazaki, & Sprott, 2000; Heavey, 1978;
Mikesell, 1989; Spiro, 1974; Suits, 1977). With the exception of the
66
FAMILY AND CONSUMER SCIENCES RESEARCH JOURNAL
Clotfelter and Cook (1989) and Mikesell (1989) studies, previous literature shows that people in lower income categories spend a greater
percentage of their income on lottery tickets and thus on the tax inherent in the price of tickets. Livernois (1987) examined both the tax and
the expenditure incidence of lottery profits in western Canada. He
concluded that the lottery redistributes income from low-income to
high-income groups.
Another area of lottery research addresses the question of who is
more likely to participate in playing the lottery and to what extent.
Province lottery agencies are very interested in this issue, as this information allows them to target these groups with lottery advertising
and lottery retail outlets. The studies by Clotfelter and Cook (1989),
Mikesell (1989), and Hansen (1995) provide a detailed examination of
the socioeconomic and demographic characteristics of lottery players
that significantly determine lottery ticket purchases in the United
States. These studies indicate that minority groups are more likely to
play than nonminority groups. They also find that lottery sales are
inversely related to educational levels, suggesting that less educated
individuals have higher lottery ticket expenditures relative to those
more educated individuals. Place of residence is also found to be a significant determinant of lottery ticket expenditures. Also, individuals
living in urban areas tend to have higher lottery expenditures relative
to individuals living in rural areas.
Kitchen and Powells (1991) examined the determinants of household lottery expenditures in the six regions of Canada using micro
data. They concluded that some household characteristics such as
wealth, age, occupation, mother tongue, and urban location vary in
the extent to which they significantly affect the level of lottery expenditures across regions. Other household characteristics such as
after-tax household income, gender, and education of the head have
the same effect on lottery expenditures regardless of the region. They
also concluded that in spite of the fact that the Canadian lottery
expenditures are found to be regressive, they are less regressive in
their impact than American lotteries.
Kitchen and Powells (1991) used the Tobit estimation technique to
obtain the parameter estimates. This technique accounts for the influence of the explanatory variables on the probability of whether to buy
lottery tickets as well as their effect on the decision regarding the
amount to purchase once the choice to buy has been made. Although
the decisions of whether to purchase lottery tickets versus how much
Abdel-Ghany, Sharpe / LOTTERY EXPENDITURES IN CANADA
67
to spend by those choosing to purchase may be influenced by many of
the same socioeconomic and demographic factors, there is no reason
to believe that an explanatory factor common to both decisions must
have the same sign and magnitude in both decisions.
Scott and Garen (1994), using a selection-bias-corrected ordinary
least squares regression method investigated individual lottery ticket
purchases in the state of Kentucky. They also estimated the lottery
ticket demand function by a consistent two-step procedure and
developed a Chow test to determine if the Tobit restrictions were
appropriate. They concluded that the use of Tobit statistical procedure in analyzing lottery purchases lead to mis-specification of the
model due to the situation of discontinuous consumer’s indifference
curves resulting perhaps from stigma and/or fixed cost of purchase.
Stranahan and Borg (1998) analyzed Scott and Garen’s model and
showed that the truncated Tobit and probit is a better methodology
for estimating lottery expenditures. Using data from telephone interview surveys in Florida, Virginia, and Colorado, their results showed
that age, race, education, and other variables have very different
effects on the probability of play versus the amount spent on lottery
products.
This study improves on previous research in several ways. First,
the Canadian expenditure data used in this study are recent. Second,
these are the only expenditure data that allow an analysis of lottery
expenditures. Third, this study uses a statistical method that offers
advantages over the Tobit analysis used in previous research (Cragg,
1971).
DATA AND METHOD
Data for this proposed study are derived from the 1996 Canadian
Family Expenditure Survey. The Family Expenditures Surveys Section, Household Surveys Division, Statistics Canada released the data
set to public use in May 1999. This survey of households was
designed to yield a representative sample of persons in private households in 17 metropolitan areas of Canada. The data set excludes persons living on Indian reservations, patients in old age homes and hospitals, persons in penal institutions, and families of official
representatives of foreign countries. Our sample from this set of data
consisted of 10,079 households. All expenditure and income quantities are annualized (Statistics Canada, 1996, p. 6).
68
FAMILY AND CONSUMER SCIENCES RESEARCH JOURNAL
Analysis of lottery expenditures presents a challenge. Not everyone would choose to participate in a lottery. A large number of zero
expenditures makes ordinary least squares analysis inappropriate
(Greene, 1993). Although Tobit analysis can account for a large number of zeros, it presumes the sign and significance of the factors affecting the decision to participate and the decision of how much to participate are the same. This presumption is not likely to be correct for
lottery expenditures. So-called double hurdle models are now established in the literature as being superior to Tobit modeling in dealing
with the problem of a large number of zero expenditures (Cragg,
1971). Unlike the Tobit model, the double hurdle model explicitly recognizes that the factors associated with the decision to participate
may have a different influence on the level of participation among
those who choose to participate. In consumption studies, double hurdle models can be used to separate the decision to consume (participate) from the level of consumption (expenditures) and, therefore,
provide more meaningful insights into consumers’ behavior than
does the Tobit model (Cragg, 1971). Moreover, although many of the
same factors (such as income and demographics) may influence both
participation and expenditure, they may have different effects on
participation.
The double hurdle model specifies a participation equation, Xα +
µ, and an expenditure (consumption) equation, Yβ + ε, such that
expenditures, E, are modeled as, E = Yβ + ε, if Xα + µ > 0 and Yβ + ε > 0; =
0 otherwise, where X and Y are vectors of explanatory variables, α
and β are vectors of parameters, and µ and ε are the error terms. As
measured in the Canadian Family Expenditures Survey, “participation” in lottery play is a “yes-no” proposition; the dependent variable
of the participation equation is a one-zero indicator. Following standard practice supported by the literature, we estimate this equation
using probit methods rather than by logit or dependent variable techniques. As expenditures on lottery are continuous above zero, the
consumption equation is estimated using truncated regression analysis (Cragg, 1971; Greene, 1993). LIMDEP version 7 (Greene, 1995) was
used to estimate the equations in this study.
RESULTS AND DISCUSSION
Variable descriptions are given in Table 1. Table 2 lists the independent variables used to specify the participation and expenditure
Abdel-Ghany, Sharpe / LOTTERY EXPENDITURES IN CANADA
69
TABLE 1: Variable Definitions Used to Examine Effect of Household Characteristics on Lottery Expenditures in Canada and Incidence of Lottery Expenditures
Variable
Definition
EX
HI
HW
Age
Urban
Average dollar expenditures per year on lottery
Household income after tax in $10,000
Household wealth (net value of dwelling in $10,000)
Age in years of reference person
1 if household lives in urban area; 0 otherwise
Education
ED1
ED2
ED3
1 if high school; 0 otherwise
1 if some college; 0 otherwise
1 if college degree; 0 otherwise
Country of birth
CB1
1 if Canada; 0 otherwise
Occupation
OC1
OC2
OC3
OC4
OC5
OC6
1 if managerial/professional and technical; 0 otherwise
1 if teaching; 0 otherwise
1 if sales/service/clerical; 0 otherwise
1 if blue collar; 0 otherwise
1 if other; 0 otherwise
1 if not stated; 0 otherwise
Type of household
TY1
TY2
TY3
TY4
1 if one-person; 0 otherwise
1 if married; 0 otherwise
1 if lone-parent; 0 otherwise
1 if other; 0 otherwise
Presence of children younger
than 15 years of age
CH15
1 if present; 0 otherwise
equations. For each variable, we show the sample average computed
over the full sample and over the subset of lottery players in each of
the six regions of Canada. Units of measurement are also given in the
table. Our choice of explanatory variables was predicated on demand
theory and was consistent with many previous studies related to this
topic (e.g., Kitchen & Powells, 1991; Stranahan & Borg, 1998).
The age, education, country of birth, and occupation indicators in
this sample of households refer to attributes of the household reference person, the household member primarily responsible for the
financial maintenance of the household. All other variables listed,
such as after-tax income, wealth, residence, type of household, and
70
TABLE 2: Variable Definitions and Mean Statistics in Examining Effect of Household Characteristics on Lottery Expenditures in
Canada and Incidence of Lottery Expenditures
Atlantic Canada
Entire
Sample
n=
2,350
EX
HI
HW
Age
Urban
Lottery
Players
n=
1,854
Quebec
Entire
Sample
n=
1,545
British Columbia
Lottery
Players
n=
1,331
Entire
Sample
n=
1,448
Lottery
Players
n=
1,157
244.11
309.42
243.93
283.15
226.83
283.88
3,541
3,762
3,435
3,525
4,045
4,206
4,462
4,605
4,195
4,274
11,250
11,390
47.4
47.8
47.8
47.7
48.2
47.9
0.72
0.730
0.89
0.89
0.93
0.93
Ontario
Entire
Sample
n=
2,396
Manitoba/
Saskatchewan
Alberta
Lottery
Players
n=
1,876
Entire
Sample
n=
821
Lottery
Players
n=
679
Entire
Sample
n=
1,519
Lottery
Players
n=
1,238
260.56
332.78
300.07
362.82
270.97
332.47
4,333
4,595
4,362
4,494
3,656
3,866
8,179
8,339
6,598
6,452
4,499
6,633
48.4
47.6
45.9
44.6
48.7
48.1
0.93
0.93
0.90
0.90
0.85
0.85
Education
ED1
ED2
ED3
CB1
0.43
0.28
0.13
0.94
0.43
0.30
0.13
0.95
0.40
0.29
0.12
0.89
0.41
0.29
0.11
0.90
0.38
0.37
0.18
0.70
0.38
0.38
0.16
0.72
0.40
0.29
0.18
0.69
0.41
0.30
0.17
0.71
0.39
0.35
0.17
0.78
0.41
0.37
0.15
0.80
0.44
0.27
0.14
0.86
0.45
0.28
0.14
0.88
Occupation
OC1
OC2
OC3
OC4
OC5
OC6
0.16
0.03
0.16
0.19
0.08
0.02
0.17
0.03
0.18
0.20
0.08
0.01
0.18
0.03
0.19
0.17
0.09
0.01
0.18
0.03
0.18
0.20
0.09
0.01
0.20
0.03
0.18
0.17
0.06
0.04
0.21
0.03
0.19
0.18
0.06
0.04
0.21
0.03
0.17
0.17
0.09
0.02
0.21
0.03
0.18
0.20
0.10
0.14
0.23
0.03
0.21
0.22
0.07
0.03
0.23
0.04
0.22
0.23
0.07
0.03
0.20
0.03
0.17
0.18
0.08
0.03
0.21
0.03
0.18
0.19
0.08
0.03
Household type
TY1
0.18
TY2
0.24
TY3
0.09
TY4
0.07
CH15
0.33
0.15
0.24
0.09
0.08
0.34
0.27
0.25
0.08
0.05
0.30
0.25
0.26
0.08
0.05
0.30
0.25
0.26
0.08
0.06
0.31
NOTE: See Table 1 for the list of definitions of the variables.
0.23
0.26
0.08
0.07
0.31
0.22
0.23
0.07
0.08
0.33
0.18
0.24
0.06
0.09
0.34
0.20
0.24
0.06
0.08
0.37
0.18
0.25
0.06
0.09
0.39
0.24
0.23
0.08
0.06
0.34
0.21
0.25
0.08
0.06
0.30
71
72
FAMILY AND CONSUMER SCIENCES RESEARCH JOURNAL
TABLE 3: Lottery Expenditures by Region in Canada
Region
Atlantic Canada
Quebec
British Columbia
Ontario
Alberta
Manitoba/Saskatchewan
% of Households
Buying Lottery
78.9
86.1
79.9
78.3
82.7
81.5
Lottery Expenditures as a
% of Household Income
0.82
0.80
0.67
0.72
0.81
0.86
presence of children younger than 15 years of age, are measured for
the entire household. For the categorical variables of education, occupation, and type of household, the following were used as reference
groups: reference person has less than high school education; reference person is not working or retired; and married couple households
with unmarried children only, married couple households with relatives only and/or with at least one nonrelative.
On a per household basis, residents of Alberta, Manitoba/Saskatchewan, and Ontario are the most prolific gamblers in terms of
Canadian dollars wagered annually ($300.07, $270.97, $260.56 for the
whole sample, and $362.82, $332.47, and $332.78 for lottery players in
the respective regions).
Table 3 shows the percentage of households that purchased lottery
tickets as well as the average expenditures of these households as a
percentage of after-tax income. The column “% of Households Buying
Lottery” in Table 3 shows that participation rates ranged from a high
of 86.1% of all households in Quebec to a low of 78.3% in Ontario. It is
interesting to note that these rates are higher than those reported by
Kitchen and Powells (1991) for the same regions a decade earlier. The
second column, “Lottery Expenditures as a % of Household Income,”
reports total lottery expenditures as a percentage of after-tax household income for all households engaging in lottery expenditures. The
percentage ranged from 0.67 for households residing in British
Columbia to 0.86 for households residing in Manitoba/Saskatchewan.
The estimates of the effects of the independent variables on the participation (purchasing or not purchasing lottery tickets) and expenditurelevel decisions of the participants on a region-by-region basis are
shown in Table 4. The results indicate that the probability of participation in the lottery is positively related to household after-tax income
in all regions, with the exception of Quebec and Alberta. Household
Abdel-Ghany, Sharpe / LOTTERY EXPENDITURES IN CANADA
73
wealth and place of residence (urban vs. rural) have no discernible
effect on whether a household purchases lottery tickets.
Participation in the lottery decreases with age in Alberta; however,
age does not significantly affect lottery participation in all remaining
regions. College graduates are less likely to be lottery players in the
regions of Quebec, British Columbia, and Ontario. Born Canadians
are more likely to participate in lottery purchases than immigrants in
the regions of British Columbia and Manitoba/Saskatchewan. The
type of occupation has a varying effect on the probability of being a
lottery player. Single persons are less likely to participate in lottery
purchases in all regions, with the exception of Alberta and Manitoba/Saskatchewan. Also, lone-parent households are less likely to
be lottery players in the regions of Quebec and Ontario. Having children younger than the age of 15 reduces the probability of the household to participate in lottery purchase in the Atlantic region as well as
Quebec and Ontario.
Table 4 also presents the estimates of the effects of the independent
variables on how much households actually spend on lottery.
After-tax household income is positively related to influence on the
level of expenditures on lottery purchases across all regions, with the
exception of Quebec. These results are consistent with the results of
the study by Kitchen and Powells (1991), in which income level was
found significantly to influence lottery expenditures in all regions.
The relationship between wealth and household lottery expenditures is positively statistically significant for households residing in
Quebec; however, the relationship is negatively related in Manitoba/Saskatchewan. Kitchen and Powells (1991) found that wealth
was negatively related to the amount spent on lottery purchases in
four of the regions, with the exception of Quebec and Alberta.
The age of the reference person is positively related to lottery
expenditures and statistically significant in British Columbia,
Ontario, and Alberta, and insignificant elsewhere. Whereas households in urban areas in British Columbia spend significantly less on
lottery purchases, households in urban Manitoba/Saskatchewan
spend significantly less than their rural counterparts on lottery
purchases.
The results broadly indicate that education has a negative effect on
lottery expenditures. More specifically, households in which the reference person had some college education spent significantly less on
lottery tickets than households in which the head had less than high
school education in all regions, with the exceptions of British
74
TABLE 4: Estimation of Participation and Level of Expenditures on Lottery in Canada
Atlantic Canada
Quebec
Lottery
British Columbia
Lottery
Ontario
Lottery
Alberta
Lottery
Manitoba/
Saskatchewan
Lottery
Lottery
Probability Expenditures Probability Expenditures Probability Expenditures Probability Expenditures Probability Expenditures Probability Expenditures
of Lottery
Among
of Lottery
Among
of Lottery
Among
of Lottery
Among
of Lottery
Among
of Lottery
Among
Participation Participants Participation Participants Participation Participants Participation Participants Participation Participants Participation Participants
n = 2,350
Intercept
HI
HW
Age
Urban
Education
ED1
ED2
ED3
CB1
n = 1,877
0.574*
225.01
(0.270) (124.31)
0.121*** 24.59**
(0.024)
(8.423)
–0.003
–4.312
(0.008)
(3.514)
–0.002
2.23
(0.003)
(1.433)
0.039
56.49
(0.071)
(32.00)
0.037
(0.090)
0.148
(0.106)
–0.266
(0.145)
0.070
(0.134)
n = 1,545
n = 1,330
n = 1,448
0.460
(0.351)
0.039
(0.028)
–0.004
(0.009)
0.006
(0.009)
0.179
(0.128)
154.10
(146.75)
20.07
(10.81)
8.06*
(3.88)
1.36
(1.69)
131.73*
(52.84)
0.264
(0.344)
0.051*
(0.022)
–0.001
(0.003)
0.005
(0.004)
0.150
(0.155)
–76.43
(49.62)
–163.40**
(57.25)
–191.66**
(76.37)
–8.06
(55.32)
–0.123
65.57
(0.161)
(70.58)
–0.163
–120.87
(0.167)
(73.31)
–0.614*** –188.92*
(0.184)
(86.23)
0.306***
81.33*
(0.086)
(40.71)
–78.54
0.165
(44.45)
(0.122)
–163.80*** 0.068
(49.40)
(0.140)
–222.19*** –0.370*
(66.27)
(0.176)
62.34
0.206
(61.46)
(0.162)
n = 1,157
n = 2,396
n = 1,882
66.09
0.683**
43.56
(156.22)
(0.263)
(158.92)
28.81**
0.078***
28.34**
(9.50)
(0.017)
(9.56)
0.174 –0.007
–3.25
(1.518) (0.004)
(2.31)
4.59** –0.001
4.68**
(1.70)
(0.003)
(1.83)
–151.46* –0.042
62.60
(69.77)
(0.131)
(72.69)
–0.057
–48.00
(0.098)
(63.08)
–0.096
–139.93*
(0.109)
(69.78)
–0.321** –264.24***
(0.126)
(81.10)
0.225***
98.10*
(0.066)
(41.16)
n = 821
n = 679
n = 1,519
n = 1,234
1.325**
(0.538)
0.044
(0.028)
–0.82E-04
(0.008)
–0.016**
(0.006)
–0.084
(0.201)
–207.42
(327.31)
60.65***
(15.77)
–8.79
(5.13)
7.03*
(3.43)
–25.46
(118.47)
–0.118
157.42
(0.328) (178.33)
0.152*** 27.35*
(0.029)
(12.46)
–0.008
–12.14**
(0.009)
(4.61)
–0.002
2.66
(0.004)
(1.98)
–0.039
187.21***
(0.118)
(58.32)
0.351
(0.199)
–0.336
(0.214)
–0.190
(0.236)
0.204
(0.128)
304.95*
(142.09)
57.25
(149.00)
–16.77
(168.93)
9.87
(84.68)
0.055
(0.120)
0.043
(0.139)
–0.340
(0.167)
0.326**
(0.105)
–99.09
(65.21)
–295.85***
(73.86)
–302.04***
(89.43)
164.36**
(60.31)
Occupation
OC1
0.220
–87.36
(0.125)
(54.12)
OC2
0.235
–80.17
(0.209)
(91.01)
OC3
0.359** –19.23
(0.114)
(48.94)
OC4
0.259*
48.00
(0.112)
(49.15)
OC5
0.184
3.67
(0.139)
(60.34)
OC6
–0.237
–68.87
(0.225) (114.10)
0.387*
(0.163)
–0.102
(0.226)
0.517***
(0.149)
0.444**
(0.62)
0.522**
(0.195)
–0.170
(0.400)
Household type
TY1
–0.506***
(0.121)
TY2
–0.171
(0.108)
TY3
0.147
(0.126)
TY4
–0.151
(0.138)
CH15
–0.315**
(0.101)
–0.599*** –1.69
(0.169)
(61.96)
–0.213
–29.89
(0.167)
(55.73)
–0.370*
–83.14
(0.164)
(69.47)
–0.366
17.51
(0.212)
(82.74)
–0.387** –76.61
(0.152)
(53.59)
–148.06**
(53.92)
–50.13
(42.91)
–102.85
(54.74)
98.57
(57.15)
–101.70**
(39.93)
–63.47
(66.06)
–75.38
(109.07)
–15.88
(58.67)
–13.44
(62.58)
61.00
(69.93)
–113.61
(192.07)
0.314*
33.97
(0.148)
(67.56)
–0.850** –45.75
(0.296) (119.20)
0.321*
–7.93
(0.140)
(62.94)
0.250
6.80
(0.152)
(67.77)
0.451*
34.59
(0.216)
(86.98)
–0.065
–119.34
(0.202) (103.11)
0.186*
22.24
(0.112)
(69.58)
–0.077
–114.64
(0.190)
(124.41)
0.227*
50.56
(0.106)
(66.81)
0.513*** 119.22
(0.119)
(66.83)
0.250
76.08
(0.132)
(46.99)
–0.153
9.70
(0.220)
(161.57)
–0.125
(0.208)
0.333
(0.363)
0.159
(0.212)
–0.030
(0.210)
0.213
(0.301)
–0.009
(0.358)
–88.32
(133.82)
–121.80
(218.68)
–11.39
(129.49)
70.87
(132.17)
–185.70
(163.51)
–215.24
(224.57)
–0.189
100.07
(0.150)
(78.38)
0.320 –101.54
(0.259) (130.62)
0.338*
6.63
(0.144)
(74.59)
0.308* 132.73
(0.149)
(76.48)
0.448* –63.36
(0.194)
(90.84)
–0.108
41.30
(0.234) (124.20)
–0.425**
(0.160)
–1.109
(0.148)
–0.078
(0.166)
–0.101
(0.203)
–0.091
(0.137)
–0.603*** –180.84**
(0.121)
(72.24)
–0.197
–116.04*
(0.111)
(60.22)
–0.447*** –97.30
(0.123)
(82.37)
–0.039
28.62
(0.138)
(74.95)
–0.282** –105.82*
(0.101)
(55.05)
–0.231
(0.220)
0.008
(0.201)
–0.297
(0.229)
0.128
(0.259)
–0.184
(0.186)
–100.94
(135.85)
25.47
(117.73)
–174.02
(150.28)
–78.35
(137.78)
–72.36
(107.14)
–0.087
(0.157)
0.161
(0.146)
0.017
(0.157)
0.146
(0.194)
–0.023
(0.135)
NOTE: See Table 1 for the list of definitions of the variables.
*p > .05. **p > .01. ***p > .001.
–29.87
(70.04)
52.97
(61.23)
–73.22
(73.02)
207.55**
(82.42)
–35.21
(56.50)
–274.44***
(80.67)
–111.72
(68.25)
–113.13
(82.08)
–77.09
(93.21)
–215.54***
(63.28)
75
76
FAMILY AND CONSUMER SCIENCES RESEARCH JOURNAL
Columbia and Alberta. In all regions with the exception of Alberta,
reference persons with college degrees also spent significantly less
than those with less than high school education. These results are
broadly consistent with previous studies (Clotfelter & Cook, 1987,
1989; Kitchen & Powells, 1991; Livernois, 1987; Scott & Garen, 1994;
Stranahan & Borg, 1998).
Born Canadians spent significantly more on lottery tickets than
immigrants in the regions of British Columbia, Ontario, and Manitoba/
Saskatchewan. The type of occupation held by the reference person
was limited in affecting the level of expenditures on lottery in all
regions. Only in Ontario, the group for which the occupation of the
reference person was designated as farming, fishing, mining, and so
forth, spent significantly more than the retired and unemployed
group.
Lone-person households spent significantly less on lottery purchases than married-couple households with unmarried children in
the Atlantic region of Canada, Ontario, and Manitoba/Saskatchewan.
Also in Ontario, married-couple households only spent significantly
less than married-couple households with unmarried children. The
presence of children younger than the age of 15 in the household has a
negative effect on the amount of money spent on lottery tickets in the
Atlantic region of Canada as well as in Ontario and Manitoba/
Saskatchewan.
The question of who bears the burden of state lotteries, or in the
language of economists, the question of the incidence of lottery
finance, has been the most frequently researched issue related to the
lottery. Some researchers argue that the revenue the state keeps from
lottery sales is a tax. Others argue to the contrary, that because there is
no coercion involved and that people voluntarily participate in the
lottery, it should not be considered as a tax (Livernois, 1986; Stranahan &
Borg, 1998; Zorn, 1988).
The regressivity of the lottery can be examined by estimating the
income elasticity of lottery expenditures at the mean. For example, a
10% increase in after-tax income for a household residing in the Atlantic region of Canada with the sample’s mean income of $37,620
increases yearly lottery expenditures from $309.42 to $318.67, which
is an increase of only 2.99%. This means that the income elasticity of
lottery expenditures at the mean is only 0.299, and an income elasticity of less than 1 denotes a regressive pattern of expenditures, which
leads to a regressive tax.
Abdel-Ghany, Sharpe / LOTTERY EXPENDITURES IN CANADA
77
The application of the same above procedure yields an income
elasticity of lottery expenditures at the mean of 0.250 (Quebec), 0.427
(British Columbia), 0.391 (Ontario), 0.751 (Alberta), and 0.318 (Manitoba/Saskatchewan). As can be noted, the income elasticity of lottery
expenditures across all regions is less than 1, denoting a regressive
pattern of expenditures, which implies a regressive tax.
CONCLUSION
Provinces currently retain 80 to 85% of gambling revenue in Canada, whereas charities share 15 to 20% of this revenue (Canadian West
Foundation, 2000). Given the tremendous growth in the lottery system in Canada since its legalization in the early 1970s, the Canadian
provinces are relying more heavily on its generated funds. Results of
this study indicate that the factors associated with the decision to participate in the lottery do not necessarily have the same influence on
the level of expenditure on lottery for those who choose to participate.
Consequently, this study provides valuable information to lottery
administrators in the different provinces of Canada for designing programs to either recruit new lottery players or to increase spending for
those who already participate in the purchasing of lottery tickets.
Whereas some household characteristics (wealth and place of residence) do not affect the probability of a particular household to participate in lottery playing in every province, others vary in the extent
to which they significantly affect such a probability. Similarly, the profile of those players who spend more on lottery playing can be drawn
from the results. In general, lottery expenditures increase as household after-tax incomes increase, and lottery expenditures decline as
the education level of the reference person increases. Other household characteristics affect the level of spending on the lottery variably
among the provinces. For those variables exhibiting regional variations, their effect should be considered before designing regional
plans.
The results also indicate that the percentage of after-tax income
spent on lottery declines as income increases in every region. Therefore, from the standpoint of public policy makers, we conclude that
the lottery tax in Canada is regressive.
78
FAMILY AND CONSUMER SCIENCES RESEARCH JOURNAL
REFERENCES
Borg, M. O., & Mason, P. M. (1988). The budgetary incidence of a lottery to support education. National Tax Journal, 61, 75-85.
Brinner, R. E., & Clotfelter, C. T. (1975). An economic appraisal of state lotteries. National
Tax Journal, 28, 395-404.
Canadian West Foundation. (2000). Canadian gambling behaviour and attitudes [Online].
Available: www.cwf.ca/pubs/200001.cfm?pub_id=200001
Clotfelter, C. T. (1979). On the regressivity of state-operated “numbers” games. National
Tax Journal, 32, 543-548.
Clotfelter, C. T., & Cook, P. J. (1987). Implicit taxation in lottery finance. National Tax
Journal, 40, 533-546.
Clotfelter, C. T., & Cook, P. J. (1989). Selling hope: State lotteries in America. Massachusetts:
Harvard University Press.
Cragg, J. G. (1971). Some statistical models for limited dependent variables with applications to the demand for durable goods. Econometrica, 39, 829-844.
Greene, W. H. (1993). Econometric analysis (2nd ed.). New York: MacMillian.
Greene, W. H. (1995). LIMDEP Version 7 [Computer software]. New York: Econometric
Software.
Hansen, A. (1995). The tax incidence of the Colorado State lottery instant game. Public
Finance Quarterly, 23, 385-398.
Hansen, A., Miyazaki, A. D., & Sprott, D. E. (2000). The tax incidence of lotteries: Evidence from five states. The Journal of Consumer Affairs, 34, 182-203.
Heavey, J. F. (1978). The incidence of state lottery taxes. Public Finance Quarterly, 6,
415-426.
Kitchen, H., & Powells, S. (1991). Lottery expenditures in Canada: Aregional analysis of
determinants and incidence. Applied Economics, 23, 1845-1852.
Livernois, J. R. (1986). The taxing game of lotteries in Canada. Canadian Public Policy, 12,
622-627.
Livernois, J. R. (1987). The redistributive effects of lotteries: Evidence from Canada.
Public Finance Quarterly, 15, 339-351.
Mikesell, J. L. (1989). A note on the changing incidence of state lottery finance. Social Science Quarterly, 70, 513-521.
North American Association of State and Provincial Lotteries. (2000a). Did you know?
[Online]. Available: www.naspl.org/faq.html
North American Association of State and Provincial Lotteries. (2000b). NASPL members–Western Canada [Online]. Available: www.naspl.org/states/wclc.html
Scott, F., & Garen, J. (1994). Probability of purchase, amount of purchase, and the demographic incidence of the lottery tax. Journal of Public Economics, 54, 121-143.
Spiro, M. H. (1974). On the tax incidence of the Pennsylvania lottery. National Tax Journal, 27, 57-61.
Statistics Canada. (1996). Microdata files [Electronic database]. Survey of Family Expenditures, 1996. Ottawa, Ontario: Household Surveys Division, Statistics Canada.
Stranahan, H. A., & Borg, M. O. (1998). Separating the decisions of lottery expenditures
and participation: A truncated approach. Public Finance Review, 26, 99-117.
Suits, D. B. (1977). Gambling taxes: Regressivity and revenue potential. National Tax
Journal, 1, 19-35.
Zorn, C. K. (1988). The lottery: Its economic effects. Indiana Business Review, 63, 4-5.