jackpot rollover and lottery regressivity

have found the lottery to be a regressive form of taxation that varies by game and whose regressivity declines at higher jackpot size. This paper conducts an in-depth analysis of the effect of consumer spending on lottery regressivity during the Mega Millions rollover sequence and reports the following findings. First, regressivity among six games examined in the paper varies by game and is inversely related to the prize/jackpot size of the game. Second, an increase in the jackpot size reduces the regressivity for the Mega Millions game, but not for the other five games. Third, the impact of household income distribution on lottery sales varies by game, and in the case of Mega Millions, by jackpot size as well. We did not find a significant difference in the demand for the Mega Millions game between below middle-income households and high-income households. However, the demand by middle and upper-middle income households is significantly higher than the demand by high-income households, especially at a higher jackpot size. Lastly, as the jackpot size grows over $100 million and higher, a large cash inflow from states with no Mega Millions flows into the New Jersey lottery market. The majority of the additional cash inflow is spent on the Mega Millions game and there does not appear to be a significant spillover to other New Jersey lottery games.