Religionist bully Jerry Cox shouldered his way on to the front page of the Democrat-Gazette today with the claim that, essentially, the Arkansas lottery doesn't take enough from gamblers. His point is that a smaller percentage of total wagering goes to beneficiaries in Arkansas — here, college scholarships — than in most other states.
But the story lacked a full discussion of some critical elements of comparing states, though Ernie P. alluded to it. The single biggest expense of any lottery is gamblers' winnings. The more you pay out (that is, the greater the odds are that a player will win), the less you have available in profits.
Payoffs to gamblers and administrative overhead as a percentage of total gambling are good ways to measure lottery performance. Here's a report on that in Arkansas for the year ending June 30. When Ernie P. talks about the dangers of changing prize structure so as to produce greater profits, he's talking about jiggering with the win percentage. You can do that and up your profit, but you can also discourage play if it's too much harder to win. (I say that even as people plug jillions into video poker, the biggest sucker bet in gambling.)
Here's a 2006 nationwide report that breaks down lottery revenue and expenses in a way to show administrative overhead and payout percentage in each state (except Arkansas, which wasn't then gambling). Unfortunately, in this form, it's not possible to compare directly with the Arkansas figures because I can't be sure what expenses are included in the various categories on the state and national reports. In direct winnings, the Arkansas average payout percentage is about that of the national figure, but there are great variances among the states and dramatically different schemes in some states (several have video lottery terminals, for example.) In short, I don't know. But I do know that administrative costs in Arkansas, even with the highly paid executive payroll, are a relatively insignificant part of the total cost picture.