Table 5. AverageVaIues of Actual Revenue (R), Optimum Revenue (OPTR), Returns to Scale Measure (u),
PureTechnicaI and Scale Inefficiency and Efficiency Ratios for Selected Farm Samples
|
Data Set |
Subset |
Obs. |
R* |
OPTR |
U |
Inefficiency |
Eff. Ratio | ||
|
Pure |
Scale |
Pure |
Total | ||||||
|
no. |
....... |
- - dollars - - |
...... |
.....dollars..... | |||||
|
Total |
170 |
183,072 |
201,951 |
0.76 |
73,772 |
25,111 |
0.71 |
0.65 | |
|
Small |
85 |
124,581 |
202,089 |
1.01 |
76,057 |
4,071 |
0.62 |
0.61 | |
|
Large |
85 |
241,563 |
201,812 |
0.58 |
71,487 |
46,151 |
0.77 |
0.67 | |
|
Small |
85 |
124,581 |
110,351 |
0.64 |
35,769 |
17,797 |
0.78 |
0.70 | |
|
Small |
42 |
101,346 |
110,310 |
0.76 |
39,189 |
7,394 |
0.72 |
0.69 | |
|
Large |
43 |
147,275 |
110,390 |
0.53 |
32,428 |
27,957 |
0.82 |
0.71 | |
|
Large |
85 |
241,563 |
243,097 |
0.81 |
52,200 |
13,459 |
0.82 |
0.79 | |
|
Small |
42 |
196,663 |
243,337 |
0.97 |
53,357 |
1,075 |
0.79 |
0.78 | |
|
Large |
43 |
285,418 |
242,862 |
0.68 |
51,071 |
25,555 |
0.85 |
0.79 | |
decreasing returns on the larger farms. This effect is
due to the form of the RHF in which the measure of
returns to scale is inversely related to output and the
fact that factor intensities do not differ appreciably
across farms (see footnote 12). Second, in situations
where factor intensities are relatively constant, the
appeal of the RHF specification examined here may
be diminished. High levels of scale inefficiencies
may be due to the specification of the RHF rather
than to the underlying nature of the production tech-
nology. In these circumstances, it may be more
useful to consider alternative parametric specifica-
tions of the production technology with emphasis on
statistical testing of the functional form prior to
efficiency measurement. Finally, the results of this
study do not indicate that mean farms always expe-
rience constant returns to scale; decreasing returns
to scale are always indicated for the average output
level.
SUMMARY AND CONCLUSIONS
This analysis examines several factors influencing
farm efficiency measurement. Farm level data for
170 homogeneous grain farms was analyzed over a
six-year period for various temporal and size aggre-
gates. The effect of temporal aggregation on farm
firm efficiency measurement was assessed using the
ray-homothetic function. The change in the decom-
position of inefficiency estimation was also ex-
plored.
The results provide some insight into the recent
mixed assessments of farm firm efficiency. The
measurement of farm efficiency appears to be time
dependent. Year-to-year events statistically influ-
119
ence efficiency measures, suggesting that policy
recommendations based on data from only one year
must to be made in a cautious manner. Multiple-year
aggregation clearly has an upward effect on farm
efficiency measurement. When efficiency is exam-
ined on a yearly basis, farms appear to be producing
between 50 and 60 percent of their potential. At
higher levels of temporal aggregation, average effi-
ciency measures increase to between 60 and 65
percent of potential. Here, temporal aggregation of
expenditure data permits a more accurate repre-
sentation of the production frontier by accounting
for irregularities caused by cash versus accrual
measurement errors and the effects of beneficial
crop rotation practices.
Overall, the results of the analysis reveal a surpris-
ingly high level of farm inefficiency over the 1982-
1987 period. Even when the study controls for
sample homogeneity and calculates efficiency
measures over larger temporal aggregates, the find-
ings suggest that output could be increased by
roughly 35 percent. The causes of this inefficiency
are not readily apparent. While differences in the
level of management are clearly affecting the find-
ings, other factors may be influencing the results.
Perhaps the majority of farmers are employing older,
less effective technologies, while more innovative
farmers have adopted more effective methods of
production. Alternatively, farmers may possess dif-
ferent objectives that may result in achieving vary-
ing degrees of efficiency.
Alternative explanations of the high degree of
inefficiency rest on the procedures used to estimate
and calculate the efficiency measures and their de-