A regression was performed using June 1998 data
for 68
Victorian local government areas.
log(PLOSS) = 0.21 + 1.47 log(PEGMs) –0.34 log(PCLUBS) – 0.27 log(PHOTELS) (Eqn
4.1)
(0.55) (15.3) (-6.2) (-3.6)
where the figures in parentheses are White’s heteroscedasticity corrected
t-statistics, R2 =
0.84, N=6016, SE = 0.21, and PLOSS stands for player losses per adult in
each area; PEGMs
is the number of poker machines per 1,000 adults; PCLUBS is the number of
club venues
with poker machines per 1,000 adults; and PHOTELS is the number of hotels
with poker
machines per 1,000 adults.
The Productivity Commission estimated an identical equation
using Queensland data for
1998. However, it obtained a better fit to the data (R2 = 0.94), suggesting
that, in
Queensland, the “model” explains a higher proportion of the regional
variation in losses per
adult. Note that the negative coefficient on venue numbers is not the result
of correlation
between machine numbers and venues.
The effect on regional gaming expenditure of an increase in EGM numbers was
evaluated
using the estimated parameters and mean values of the explanatory variables
(as described
by Equation 3.6 in Section 3.2).
Back to gambling
news archive
|
