Methodology

The ETI purchasing power expenditure estimates are based on detailed analyses of consumer expenditures by the U.S. Census Bureau Consumer Expenditure Surveys. Retail purchasing data are taken from the most recently released (December 2003) Consumer Expenditure Survey for the four quarters of the year 2002. Sixteen categories of expenditures are analyzed: food at home; food away from home; apparel and related services; television equipment, tapes and discs; audio equipment, CDs, and tapes; household textiles; furniture; floor coverings; major appliances; small appliances and housewares; computer hardware and software; miscellaneous household equipment; non-prescription drugs and supplies; housekeeping supplies; personal products; and home repair commodities.

Most expenditure data are calculated from the larger CEX interview file, which contains more than 30,000 records with complete income and expenditure responses for the year. The three categories of expenditures not in the interview file (non-prescription drugs, housekeeping and cleaning supplies, and personal care products) are drawn from two years (2001 and 2002) of the CEX diary file, which includes more than 22,000 respondents with complete income and expenditure data.

Expenditure patterns are then combined by income levels and family/household types with Wisconsin Department of Revenue 2002 income tax filing data, controlling for a similar matrix of income levels and family/household types. State income tax data for non-dependent, non-elderly filers are obtained by geographic level to construct final estimates. Data on elderly households are drawn from the 2000 U.S. Census. Listings of expenditures included in the 16 categories analyzed are detailed in the methodology section of each report.

The Milwaukee purchasing power profiles differ from those available on the Employment and Training Institute website. The Milwaukee profiles use neighborhood population/income data derived primarily from the most recent State of Wisconsin income tax filing data (tax forms filed in 2003 for 2002 tax purposes), reflecting recent economic changes by neighborhood rather than 2000 U.S. Census survey data on income. Secondly, the Milwaukee ZIP profiles are based on actual ZIP code boundaries, rather than Census Bureau ZCTA (ZIP Code Tabulation Area) boundaries (aggregations of census blocks only approximating ZIP code areas).