arbitrary weight intervals greatly improve the quality
of information available to livestock producers for
making management decisions, they may not
necessarily reflect patterns inherent in the pricing of
beef cattle.
ANALYSIS OF LOUISIANA CATTLE PRICES
A research project was initiated in August, 1972,
to evaluate calf and yearling prices and price patterns
in Louisiana. The study was designed to determine
what modifications, if any, were desirable in price
reporting classifications to provide beef producers
with adequate price data for sound production
management decisions.
Methods and Procedures.
Weight, price and sex data were secured from one
representative auction located in each of four areas of
Louisiana: Northeast, Northwest, Central and
Southwest. Data were obtained from all regular
weekly sales of calves and yearlings within the
150-800-pound liveweight range for the two-year
period June 28, 1970, through June 27, 1972. The
sample included 158,192 observations (animal sales)
consisting of 73,937 steers, 41,655 heifers, and
42,600 head upon which sex identification could not
be determined. Animals are not graded at Louisiana
auction markets; therefore, information on grades
was not recorded.
Analyses were made for each of the four auction
locations separately and for the four auctions
combined. The relationships found for the individual
auctions indicated the data could be treated as single
population.
Simple Hnear regression equations were fitted to
the combined data to determine the effect of weight
on price for all animals, for steers, and for heifers.
Data for the two-year period were combined by
weeks for the determination of seasonal price
patterns. Animal weights were categorized into 26
groups of 25-pound intervals each. The Dummy
Variable Method, which accounted for the trend
during the two-year period, was used to compute the
seasonal pattern of prices for all animals for each of
the 26 weight groups. Regression model (1) was used
to calculate the seasonal price patterns.
52
(1) P = bT+ Σ aiDi
i=ι
where:
P = average weekly prices,
b = coefficient of regression for trend,
T = time (weeks, 1,2,..., 104),
aɪ = regression coefficient for the dummy
variable, and
Dj = dummy variable (1 for the ith week in the
year, 0 otherwise).
Similarity of seasonal price patterns for
successive weight groups [equation (1)] were tested
by equation (2).
к
(∑d2t- Σ ∑df)∕(k-l)
(2) F = ------------------
к к
Σ ∑d∙∕∑ ni-k
i=ι i=ι
к
d.f. = (k-l); ( ∑ni-k)
i=ι
where:
∑d21 = residual sum of squares for total
regression,
k 2
.Σ ∑dj = sum of individual sums of
1-1 squares,
к = number of weight groups, and
nɪ = number of observations in the ith
group (i= 1,2, ...,k).
Successive weight groups were combined if their
seasonal price patterns were not significantly
different at the .05 probability level. The six weight
groups which resulted from this analysis of the
combined data for all animals were used for further
analyses.1 Equation (1) was used to compute the
seasonal pattern of prices for all animals within each
of the six weight groups. Analysis of со-variance was
used to determine the effects on price of animal
weight, auction location, sex of animal, weight group
and the double and triple interactions of these
variables. Only animals identified by sex were
included in the analysis of co-variance.
Results and Interpretation.
The simple Hnear regression of weight on price
ɪ This procedure resulted in four weight groups for animals between 1 SO and 600 pounds and three weight groups for
animals between 601 and 800 pounds. A comparison of these weight groupings with those obtained from similar analyses
conducted for each of the four auction markets separately (using all observations and with due recognition of missing data)
suggested that the three weight groups from 601 to 800 pounds could be combined into two weight groups. Seasonal price
patterns for the groups combined in this manner were not significantly different at the .06 probability level.
160