Small area demand and supply estimation accounting for spatial interaction patterns: The mismatch of tobacco sales and customer base within the census tracts of Dallas County, Texas

The objective of this study is to develop a methodology that estimates and subsequently models the discrepancy between the small area tobacco supply and demand in Dallas County, Texas, on a census tract level for the years 2005 to 2009.;Small-area statistics of the tobacco supply on a census tract level are not obtainable from governmental sources despite detailed records on the taxation of tobacco products by local outlet. The supply, therefore, is estimated from secondary data by using a geo-spatial algorithm to delineate the sales areas, weighted by a proxy for the sales volume, while accounting for competing sales at other locations. On the other hand, the local demand for tobacco products is not available and must be estimated by combining the characteristics of the local populations with global reference statistics and a day/nighttime population adjustment.;Novel and efficient methodologies to obtain these estimates are presented. These estimates not only allow for monitoring the dynamics and discrepancies of the tobacco supply and demand at the local level, but they also bear the potential of uncovering potential covariates relating to the local tobacco over- or under-supply that often are masked by national or state statistics. These estimates also will benefit the Texas Comptroller's Office to identify the potential locations with tobacco sales from out-of-state sources.