Wilson
Hedonic Prices
TABLE 2. Tests of Significance of Classifi-
cation Variables on Malting Barley
Prices (F Ratio).
Clas- |
1978/79 |
1979/80 |
1980/81 |
1981/82 |
Month |
244.98* |
51.11* |
33.05* |
15.16* |
Variety |
0.19 |
1.78 |
0.90 |
14.38* |
Grade |
0.29 |
0.45 |
0.07 |
2.77 |
* Indicates rejection of the null hypothesis at the 5 per-
cent level of significance.
plicit prices for both plumpness and pro-
tein were not significantly different across
grades, 2) implicit prices for protein were
not significantly different across varieties
except in 1979/80, and 3) implicit prices
for plumpness were significantly different
across varieties except in 1981/82.
Variety, grade, and month (because of
the pooling) were included as intercept
shifters in the empirical model. Statistical
tests were used to determine whether the
effect of these classification variables on
the price of malting barley were statisti-
cally significant. The null hypothesis being
tested is that the estimated coefficients of
the classification variables are jointly equal ■
to zero. Rejection of the null hypothesis
indicates that significant differences exist
in malting barley prices which are related
to that classification variable. The results
of these tests are shown in Table 2. The
hypothesis of equality of coefficients across
months was rejected in all years, indicat-
ing that this classification variable was sig-
nificant. The effect of variety was insig-
nificant in all years except 1981/82. The
hypothesis that the intercept shifters re-
lated to grade are equal, could not be re-
jected in any of the years. This indicates
that given the covariates in the model,
price differentials could not be attributed
to the sample’s grade.
The empirical model (equation 5) was
also tested for constancy of the marginal
implicit prices by including second and
third order polynomials in plumpness and
protein. The results yielded insignificant
second and third order parameters in pro-
tein, and an insignificant third order pa-
rameter in plumpness. Estimates of the
empirical model reported here incorpo-
rate the results of hypotheses posed and
tested above. In particular, grade was not
included as an intercept or slope shifter,
and a second order parameter was includ-
ed for plumpness.
In addition, restrictions were placed on
values of the slope coefficients for plump-
ness and protein across varieties according
to the results of the tests of hypotheses.
These restrictions were different for each
year and are as follows (βla, β2a, and ‰
refer to the slope coefficient for protein,
plumpness, and plumpness squared, re-
spectively, where a denotes variety):
1978/79: β11 = β12, β21 ≠ β22, β31 ≠ β32
• 1979/80: βll ≠ β12, β21 ≠ β22, ‰ ≠ β32
1980/81: β11 = βl2 = β13 = βli, β21 ≠
β22 ≠ β23 ≠ β24, ‰ ≠ ‰ ≠
βs3 ≠ βsA
1981/82: βll = β12 = β13 — β14, β2l =
β22 = β23 = ^24> βsι ~ β32 ~
β33 = βs4-
The estimated coefficients for the he-
donic price functions are presented in Ta-
ble 3 for each crop year.10 The binary
variable for variety represents the inher-
ent value of a variety relative to Beacon.
In the first three years of the study, the
statistical results indicated that this clas-
sification was insignificant. In 1981/82,
however, varieties had statistically signif-
icant differences in their inherent value.
The coefficients'indicate that the inherent
value of Morex was 12c per bushel greater
than Beacon, but those for barker and
Glenn were not significantly different than
Beacon. Prior to 1981/82 barker was the
industry standard, Glenn and Beacon were
10 The hedonic price functions were also estimated
with the above assumptions relaxed and without
the inclusion of “Feed Barley.” These results are
very similar to those reported here (see Wilson and
Crabtree, pp. 19-22).
35