Imagine a politician proposing a tax on blacks and other minorities under the ruse of increasing funding for education in their state. This post reveals nothing new; the topic has been researched in other states many times. I just want to further demonstrate the racist nature of lotteries using North Carolina's Education Lottery as the most recent example. (I'm not against lotteries; I'm against states giving themselves the monopoly right to be the sole purveyor of lottery games to the detriment of the least advantaged and the least educated.)
Using data for North Carolina counties I regressed the log of per capita lottery spending on per capita federal dollars spent in each county (pcfedspend [in thousands]), the percentage of blacks residing in each county (black), the percentage of hispanics residing in each county (hispanic), the percentage of residents in each county with a BA degree or higher (BA), and per capita incomes for residents of each county (percapinc).
pcfedspend .0265397 (.0098) P=0.008
black .0178301 (.0022) P=0.000
hispanic .0277127 (0070) P=0.000
ba -.0160224 (.0050) P=0.002
percapinc .0000423 (.00001) P=0.014
_cons 4.439951 (.09586) P=0.000
F(5, 94) = 20.5
R-sq = .5531
What this says is that for each $1,000 increase in per capita federal spending in a county over the mean, per capita lottery spending increased by 2.65%, or about $7. (Note that North Carolina has two major military facilities. Some indeed comes from that military spending, but some certainly comes from federal welfare spending. This confirms Sam Papenfuss's argument that states adopt lotteries as a means of taxing federal welfare benefits.)
For each one percentage increase in the black population of a county, per capita spending on lottery tickets increased by 1.78%, or $5 per capita. (Note: Numerous studies have shown that blacks are no more likely to play the lottery than whites, but those that do play spend far more on lottery tickets than whites that play.) Also note that the range for black residents is one-half percent black (Graham County) to 61.7% (Bertie County) of a county's population that is black. The average is about 21% (Wake, Cleveland and a few more are about this), which means that during the first two years of the NC Lottery's existence, per capita lottery spending in Bertie County is estimated to be about $200 more than Wake County.
Each percentage point increase in a county's hispanic population increased per capita lottery spending by about 2.8%, or $7 per capita.
And for each percentage point increase in a county's population with a bachelor's degree (ba), spending on lottery tickets decreased by about 1.6%, or $3.60 per capita.
Note that I've controlled for income (percapinc), so any increase in black and hispanic spending on lottery tickets is not a function of income, but (at least partly) of race and ethnicity. Note also that my results show essentially no effect from income on lottery purchases, which counters previous studies. My data covers only the first two years of the lottery's existence. It is likely that since it is new it has drawn people from various income levels. This is likely to change over time as the novelty quickly wears off and reason kicks in.