3
Equivalence Scales
Once a dollar figure for a reference family threshold is adopted, the next question is how to set thresholds for families of different sizes and compositions. It is widely accepted, for example, that the amount of money a four-person family requires to meet basic needs is different, and higher, than the amount needed by a single individual. An equivalence scale can be used to adjust the reference family threshold to represent equivalent amounts of money for different family types to avoid poverty.
The current poverty measure contains only an implicit equivalence scale. It does not lay out principles by which thresholds should vary across family types. Rather, Mollie Orshansky constructed the original poverty thresholds by pricing out the U.S. Department of Agriculture Economy Food Plan for different families. She developed separate food budgets for families based on the sex of the family head, family size, number of children, and, for one- and two-person units, the age of the head. For families with three or more people, these budgets were then multiplied by three to account for other basic needs. Thresholds for one- and two-person families were calculated separately based on observations of their consumption patterns that indicated smaller economies of scale (National Research Council, 1995:162-163). Thresholds were also set at a lower level for farm families than other families—a distinction that was dropped in 1981. There are currently 48 thresholds for families of different sizes and compositions in the official poverty measure.
One problem with the current implicit equivalence scale is that it takes into account only economies of scale related to food but not other items, such as shelter. It also contains a number of irregularities. For example, while economies of scale are thought to increase as families get larger, this is often not the case with the current scale, in which the addition of a fourth person adds considerably more dollars to the threshold than, say, the second person—a counterintuitive feature. Recognizing these problems, the National Research Council (NRC) report offered a set of recommendations for improving the equivalence scale.
The 1995 NRC report acknowledged that the adoption of any particular equivalence scale requires judgment. After reviewing a number of options, the report recommended a scale that took two factors into account: (1) children consume less on average than adults and (2) there are economies of scale in households so that a decreasing dollar amount should be added to the poverty threshold for each additional family member. Again, the thinking behind the latter feature is that adding a second adult to a family should raise the threshold by a higher dollar amount than, say, adding a fifth. Mathematically, the recommended scale takes the following form:
equivalence scale = (A + P*C)F,
where A equals the number of adults in a family, C equals the number of children, P is a parameter describing the proportion of the cost for an adult that a child should cost, and F is a parameter describing the extent of economies of scale. If P equals 1, for example, then children are assumed to consume the same amount as adults. If F equals 1, then no economies of scale are assumed, as each additional adult adds the same dollar amount as the previous adult. The panel recommended that P should equal 0.70 (children are assumed to consume seven-tenths of the amount consumed by an adult), and F should be set between 0.65 and 0.75. Census Bureau reports on experimental poverty measures often used the midpoint of these F values (0.70) (Short et al., 1999; Short, 2001a).
David Betson (University of Notre Dame), a member of the panel that authored the NRC report who has written extensively about equivalence scales, reiterated his views that a reasonable scale should be guided by the assumption that the marginal cost of adding an adult or child should decrease with an increase in the number of adults and that children should cost less than adults (Betson, 2004), with one exception. The exception is
based on research on the cost of children: the threshold of a single parent with a child should be roughly similar to one of a childless couple because shelter costs for a single parent with a child may exceed those of a childless couple, even if other costs (e.g., food) are lower for the former.
Thus, Betson offered an alternative “three-parameter” scale (see Betson, 1996), to be applied to the reference family threshold, consisting of food, clothing, shelter, and utilities. The scale is similar to the NRC-recommended scale, except that the third parameter provides more similarity between the expenditures of families that consist of one parent and a child and those of childless couples, and it also has larger economies of scale between single people and childless couples. Mathematically, the three-parameter scale is defined as follows:
single individual: 1.00
childless couple only: 1.41
single-parent families: (A + α + P*(C – 1)) F
all other families: (A + P*C) F,
where α = 0.8, P = 0.5, and F = 0.7. As before, A is the number of adults in a family, and C is the number of children.
Many workshop participants voiced support for the three-parameter scale to set thresholds for different family types.1 Timothy Smeeding (Syracuse University) noted: “It seems to me that this three-parameter scale that David [Betson] has worked on produces sensible results….” The participants also voiced support for continued research on equivalence scales. Topics for future research include the effect of alternative equivalence scales on poverty rates over time, whether there are changes in economies of scale over time, the appropriateness of the parameter values adopted, and whether more factors should be taken into account in equivalence scales in the future, such as the ages of children and the value of household production by stay-at-home parents.