Per Capita Sales of Tobacco

Data Description:

These data report per capita bimonthly sales data (packs/month) for the adult population (18 years and older) for the 50 states, the District of Columbia, and the United States as a whole, for the period December 1983 through January 2003. Two levels of smoothed data are included: raw bimonthly mean, and SABL trend data. The data are a composite of monthly sales data and population data from the US Census. Until it was disbanded in late 1998, this data was collected by the Tobacco Institute for federal tax reporting purposes. Since then, the same group responsible for compiling the earlier data has been producing it through the economic consulting firm of Orzechowski and Walker, with support from the tobacco industry.

The original sales data reports monthly sales; this data is from wholesale warehouse removal data, so there is variation that likely reflects business practice, rather than actual consumption (e.g., levels of removals in the last month of any quarter is strongly correlated with the removals in the first month of the next quarter). To remove this source of variability, data were combined into 2-month intervals starting with December/January.

Sales values were then divided by population estimates to calculate per capita sales. Population estimates were done using a linear regression (by state, and for the total US), using yearly Census population numbers to interpolate monthly population values. This was divided into the bimonthly data to produce the raw bimonthly data reported in the data.

Variables: (values are per capita packs/month sales)

1. State (character) - 2 character postal abbreviations (US=United States, CA=California, etc.)
2. Year (numeric) - Year of data
3. Months (character) - Months averaged to provide data (February/March, etc.); December/January data is for the December of the year given, and January of the next year (e.g., 1989 December/January = Dec, 89 and Jan, 90)
4. Bimonthly Mean (numeric) - Packs per adult
5. SABL trend (numeric) - Packs per adult; a special smoothing procedure (SABL procedure from S-Plus statistical package) was used to separate real changes in consumption from changes due to seasonal variations

Data Files:

1. ASCII CSV
   - Max rec len: 71
   - Min rec len: 36
   - Total number of records: 6293
2. Excel spreadsheet

Caveats:

Self-reported consumption data generally underestimates sales data. This may be due to a number of causes: some sales may go out of state; not all cigarettes purchased are actually smoked; smokers may round daily consumption down to the nearest half-pack.

Per capita numbers are calculated for the total adult (18 years and over) population; no correction is made for the smoking rates, or for sales or smoking rate of the adolescent (under 18) population.

See Chapter 1 of Tobacco Control Successes in California: A Focus on Young People, Results from the California Tobacco Surveys, 1990-2002 for further discussion.

Sources:

Becker RA, Chambers JM, Wilks AR. The New S Language: A Programming Environment for Data Analysis and Graphics. Pacific Grove, CA: Wadsorth & Brooks/Cole; 1988. (SABL procedure).

Orzechowski & Walker. The Tax Burden on Tobacco. Historical Compilation, Vol. 37, 2002. Arlington, VA: Orzechowski & Walker; 2003.

The Tobacco Institute. Monthly State Cigarette Tax Reports. February 1983 through May 1998. 1875 I Street, N.W. Washington, DC 20006.

US Bureau of Census. Resident population for Selected Age Groups: 1980-1989. US Bureau of the Census, Population Distribution Branch; 1990.

US Bureau of Census. Resident population for Selected Age Groups: 1990-1996. US Bureau of the Census, Population Distribution Branch; 1996.