Demographics and Groceries

Yesterday my family threw a party for my 30th birthday. At age 80, my grandmother was present both in mind a body, which I find impressive. During dinner, we discussed how trips to the grocery store became more frequent over the years. Apparently one visit a week was the norm in the 50s and 60s (Thursdays evenings), whereas today I keep less than a day of inventory in my cupboard. How could this be?

I first thought that perhaps the cost of going to the store was higher (maybe gas was more expensive). But what caused the changed in habit is family size: I live alone, whereas my grandmother had a husband and three children. Assuming that the utility of a given meal type varies randomly in time (today I want chicken, but I can't predict what I will crave tomorrow), the advantages of preselecting a meal becomes diluted as the number of mouths who will be fed increases. In fact, when one is alone and choses his meal a few minutes in advance by going to the grocery store right before dinner, the cost of imperfect selection is always zero: you choose what you want. Whereas when a couple goes to the grocery store right before dinner, one person get to eat what he wants, and the other will have to tag along for a sub-optimal meal. Thus, given a large family size and a single meal, it becomes impossible to tame the cost of imperfect menu selection. Buying the food just-in-time only helps out one person.

By analyzing the distribution over time of meal preference, it is possible to build a pleasure cost model for couples. To lower the cost, you need to have overlapping preferences, and the distribution should be as flat as possible over time. Otherwise you may have to cook two meals at once.