Project PCW Notes
December 31, 2018

Gravitec Development is undertaking a major exploration of supermarket per-capita expenditures.  This email is another set of notes that discuss our findings.

Our last release examined Supermarket PCW and the way it is influenced by living in one of the major metro areas of the country, called a “PSU” by the BLS.  Not too surprisingly, we saw that the PCW was highest in Anchorage, Alaska, even when controlled for income and household size.

This time we will look at the influence of Race on the Supermarket PCW.  The table below shows the PCW for households that are White, Black, Asian and Hispanic.  Keep in mind that White, Black and Asian are mutually exclusive; a respondent cannot belong to more than one category.  However, Hispanic is NOT mutually exclusive to the other categories.  One can be White Hispanic, Black Hispanic, or Asian Hispanic (Robert Moon was an analyst for Kroger and was Asian Hispanic).

All of the PCW figures below are controlled for the differences in income and household size.

The Actual PCW was the PCW actually observed in the group data.

The Calculated PCW was the expected PCW calculated using just the group household size and income.

The Coefficients show the difference between the Actual and Calculated PCW.  For example, if a household declares itself to be Hispanic, then a PCW calculated with household size and income must also apply a coefficient of .879 to be more accurate.

To further show the effects of being Hispanic, the results for White-Hispanic and White-nonHispanic are also shown below.  The differences between the two groups support our .879 “Hispanic” factor.