SOUTHERN JOURNAL OF AGRICULTURAL ECONOMICS
JULY, 1983
DISCUSSION: ASSESSING STRUCTURAL CHANGE
IN THE DEMAND FOR FOOD COMMODITIES
Michael K. Wohlgenant
The topic of structural change in demand for food is
so broad that it is not surprising that Haidacher chose
to focus on the conceptual basis of structural change
rather than specific causes of structural change in de-
mand for food. Haidacher first defines what he means
by demand structure. Second, he gives two examples
on how to implement this framework empirically.
Third, he discusses problems of assessing changes in
demand structure, emphasizing the intractabilty of ob-
taining direct evidence on structural change. Finally,
he proposes and illustrates what he calls an “indirect”
approach to assessment of structural change. I will first
present a brief overview of the concepts of demand
structure and structural change, derived from Hai-
dacher’s paper and my own assessment. Then I will
comment on specific points raised in the paper relating
to the assessment of structural change. Finally, I will
conclude with some remarks on the usefulness of
household production theory as a framework for as-
sessing structural change in demand for food.
A BRIEF REVIEW
Haidacher defines the demand structure as the set of
parameters and the form of the functions that are
uniquely specified by the utility function. I would add
to this determining factor the factors of the nature of
the opportunity sets facing consumers and the com-
position of the population of consumers. The reason for
including opportunity sets in this definition is that in
some instances the budget constraint may not be lin-
ear. A situation in which this occurs is when the house-
hold is both a producer and consumer of the
commodities, producing basic goods (the direct ob-
jects of consumer choice) with market goods and other
inputs (such as household time) through household-
production functions. These production functions need
not exhibit constant returns to scale. This means the
budget constraint for the outputs of household produc-
tion may be nonlinear and concave to the origin. This
is the so-called household production model. Other
situations that can give rise to nonlinearities and kinks
in the budget constraint are discussed by Deaton and
Muellbauer (chapter 1). Whatever the source of the
nonlinearity, the nature of the opportunity set and
budget constraint—and factors affecting them—are part
of the demand structure.
Composition of the population can also be an im-
portant dimension of the structure. Haidacher only
mentions this aggregation problem in passing. When
working with aggregate data, we should not forget that
the general restrictions of consumer behavior—homo-
geneity, symmetry, and adding-up, only hold strictly
for individuals, and that market demand—even when
formulated in terms of a “representative” con-
sumer—still depends on the income distribution and
other characteristics of the consuming population.
In light of the above definition of structure, one view
of structural change might be any change in the utility
function, opportunity set, or composition of the con-
suming population. To the extent that these changes are
not accounted for by theory—through relative price and
income changes—this definition seems logical. How-
ever, as pointed out by Haidacher, this definition is
really intractable. This is because if our maintained
hypothesis (MH) is no structural change, and we reject
MH, this does not necessarily mean we accept the hy-
pothesis of structural change. This is because our al-
ternative hypothesis (AH) is unspecified. It may be
structural change or it may be specification bias of one
variety or another. Since we have no objective criteria
upon which to make a selection, the choice is arbi-
trary. Thus, he concludes that direct evidence of struc-
tural change is intractable and that we should seek other
altenatives. Haidacher then proposes an indirect ap-
proach to assessment of structural change that includes
using the conceptual framework of a complete demand
system to estimate the demand parameters, validating
the estimated structure, and indirectly assessing the re-
sults and performance of the model for possible struc-
tural change. Finally, Haidacher suggests for practical
solutions to the problem of structural change: (1) in-
cluding trend variables in the demand equations (in-
tercepts in log differential equations), (2) extending the
validation phase to sample observations outside the
period of estimate the structure, (3) incorporating dy-
namic aspects in the basic demand structure, and (4)
using econometric methods that take account of con-
temporary developments on time-variant parameters,
that is, varying parameter estimation procedures.
Michael K. Wohlgenant is an associate professor of agricultural economics, Texas A&M University.
Invited discussion presented at the annual meeting of the Southern Agricultural Economics Association, Feb. 6-9, 1983 Atlanta. Invited discussions are routinely published in the July SJAE
without editorial council review but with review by the copy editor (as per Executive Committee action June 25, 1982),
39