For many years the Utah State Tax Commission has produced detailed statistical reports based on federal tax returns filed by residents of the state. These reports have always been return-based, by which we mean that the primary unit of analysis was the tax return. In other words, tax returns were grouped according to the adjusted gross income (AGI) reported on the return. One problem with this approach, for example, was that minors working part-time were not grouped with their families and thus, our statistics always showed a high number in the low-income classes.

This report groups returns into larger units, such as households or families. What we have done is group returns that show the same last name, ZIP Code, and first part of the address into a unit, generally a family. What this will do is group minors with their parents, if they use the same name and address. It would also group a married couple living with the husband's parents (same last name). However, it would not group a married couple living with the wife's parents (different last names).

We do not think the results are without problem, but we do think it is useful to attempt a household approach, even if the results are only suggestive. (Please see previous reports for more methodological detail.) One indicator of the success of this effort would be to compare it to Census figures. Unfortunately the latest Census estimates available are for July 1, 1995. The date for our figures is ambiguous, but the first part of 1996 is a good estimate. Whereas we estimate roughly 641,583 households, the Census estimates fewer, about 617,000. They also have a larger household size: 3.12 compared to our 2.85. The Census thus estimates a larger population than the number included on tax returns, which is not surprising, since all residents do not have to file returns. Thus our method of collapsing returns into households is not perfect, as indicated above, but is better than using raw returns.