Project PCW Notes
January 9, 2019
Gravitec Development is undertaking a major exploration of supermarket per-capita expenditures. This email is another set of notes that discuss our findings.
Last week we examined Supermarket PCW and the way it is influenced by major metropolitan areas (PSUs). Not surprisingly, we noted that Anchorage has the highest PCW when controlling for household size and income.
This time we will look briefly at an ancillary issue: alcohol sales. It is easiest to understand by separating alcohol purchases into 1) beer and wine, and 2) liquor. It is also more relevant to examine household alcohol purchases rather than per-capita alcohol purchases, because we get better predictability if we assume only the first two people within the household will be of drinking age. That’s not always true, but it is more often true than not.
The following table summarizes alcohol purchases per household based on household size.
| Household | Household | |
| HH Size | Beer Wine | Liquor |
| 1 | 3.09 | 0.64 |
| 2 | 3.12 | 0.73 |
| 3 | 2.12 | 0.19 |
| 4 | 2.69 | 0.17 |
| 5 | 2.15 | 0.06 |
| 6 | 2.17 | 0.04 |
| 7 | 0.77 | 0.02 |
The following graph shows the data in more visible terms. As you can easily see, the relationship between household size and beer/wine purchases is quite strong, but the relationship to liquor purchases is hardly there at all. As far as buying beer and wine, it can be said that the more kids you have at home, the less time (or money) you have for drinking, while the inclination may be debatable.
When looking at the relationships of alcohol purchases to household income, once again we find that beer and wine purchases are more easily predicted, while the liquor consumption is relatively flat. 