4
and enumerators, or miscalibration of measur-
ing devices. A certain degree of measurement
error is inevitable, but some indicators may
face more serious measurement problems than
others. Household income data, for instance, is
criticized for its high degree of unreliability,
relative to household expenditure data.
Conceptual inequivalence and measurement
errors, in fact, are often trade-offs. The more
closely related an indicator is to its underlying
variable, the more difficult and costly its mea-
surement tends to be. This, of course, is the
rationale behind using proxy indicators. For
instance, "quantities of nutrients consumed"
would be conceptually closest to what research-
ers are interested in when measuring food con-
sumption, yet "number of meals eaten" is often
used instead as a proxy because it is easier to
measure.
The appropriateness of ways in which data
are analyzed and interpreted is also a matter of
concern. For instance, the usefulness of income
elasticity of food consumption (expenditures or
intakes) measures, as applied in a number of
studies as a measure of access-consumption
linkages, is questionable. One reason is that
elasticity estimates for household samples can
vary widely depending simply on the size and
socioeconomic characteristics of the samples
chosen. As a result, valid comparisons among
data sets, or generalizations of findings, are not
possible unless specific information identifying
a household's income level, landholding size,
place of residence (especially urban versus ru-
ral), or other factors that explain the varying
relationship between income and consumption
is available and controlled for. At the very least,
the initial income or calorie adequacy levels of
households need to be known and accounted
for before meaningful interhousehold or
intersample comparisons regarding expenditure
habits and consumption linkages can be in-
ferred from elasticity estimates.
Aggegating and averaging data is also a
problem. Often, "elasticity" studies draw infer-
ences from comparisons of elasticities estimated
from mean levels of income, caloric intake,
farm size, etc., over aggregated (either totally,
or according to income groups or other divi-
sions). An obvious limitation of this approach
is that they tell little about those at the lowest
income (or food consumption) levels. This is
true (though less so) even if households are
broken down into smaller income subgroups
(e.g., income quintiles). In fact, it would not be
surprising to find an income elasticity of calorie
consumption at the median income level of a
group (or subgroup) of households to be nearly
zero, while the elasticity for the poorest house-
holds might be nearly one.
An even more important criticism of elas-
ticity of food demand estimates is that the re-
sponsiveness of food intake to changes in in-
come, and the responsiveness of food adequacy
to changes in income, are not the same
(Ravallion 1990; Anand and Ravallion 1993).
For example, a low income elasticity of nutri-
ent intake does not necessarily imply that ag-
gregate undernutrition (as measured by a
"headcount" index) is unresponsive to income.
This distinction between the responsiveness of
food intakes and food adequacy to income
changes would be especially evident in cases
where a large proportion of the sample popula-
tion is consuming food at or near the minimum
requirement levels.
Definitions of Key Terms
Analyses of the nature and extent of linkages
among food availability, food access, food con-
sumption, and nutritional status may depend
critically on how the variables are defined (Schiff
and Valdes 1990a). Therefore, since these vari-
ables have been defined in various ways in the
literature, it is important to define them here
explicitly in order to avoid ambiguity.
Food availability, in this report, refers to
the supply of food in a nation, region, or local-
ity. Sources of supply may include home pro-
duction for consumption, domestic commercial