WHO PAYS FOR THE MARYLAND LOTTERY?: EVIDENCE FROM POINT OF SALE DATA

Abstract

For 42 states and the District of Columbia, lottery games provide a steady revenue source which can bring in up to 4% of annual state revenues. Unlike other public finance tools such as income taxes, lottery products enjoy substantial popularity. A particular reason they appear to be popular among state policymakers is that lotteries are often referred to as a “voluntary tax.” Compared to income taxes, however, relatively little is known about who plays the lottery and thus who provides states with this stream of “voluntary tax revenues.” State government budgets in the US are currently under pressure for both structural and cyclical reasons. As a result, many are debating the expansion of state-sponsored gambling to enhance their fiscal revenue streams. Previous studies have presented evidence that lotteries are regressive. For the current public policy debate to be fully informed, it is important to have information about who pays the “voluntary tax.” This paper uses unpublished FY 2005 sales data, provided by the Maryland Lottery Commission, from every individual lottery point of sale terminal in the state of Maryland. Because our data include the physical addresses of the terminals, we use an interdisciplinary approach to analyze this data by incorporating some tools from contemporary economic geography into our research. First, we marry the lottery-terminal data to census tract data and examine it using geographic information system (GIS) maps. We use the GIS maps to explore the relationship between race, income, per-capita lottery sales and the density of lottery retail sites. We also use the maps to attempt to shed some light on how different racial and income groups contribute to total state lottery revenues. To the best of our knowledge, this paper is the first in-depth use of GIS techniques to explore these questions.Because the data include not only the physical address of the terminal, but sales data disaggregated by lottery game, we are able to show the geographic distribution of the sales of particular lottery games, offering insight into how different games are played by the populations in different areas of the state. These analyses reflect new directions in the study of the lottery as a public policy issue, offering insights beyond the usual research on income and demographic effects.Finally, we present the results of a regression analysis of the effect of various social and economic factors on per-capita lottery sales. Our results confirm the findings of many previous studies, but also demonstrate that the nature of the demographic “burden” of the lottery may not be the same in Maryland as it is in other states where studies have been conducted.We are grateful for the assistance of Joe School and Anthony Dowell, for research support provided by the Special Research Initiative Fund at UMBC, to the Maryland Lottery Commission, and to participants of the Symposium on Gambling, Prediction Markets, and Public Policy hosted by Nottingham-Trent University. The opinions expressed in this paper are solely those of the authors.