ing to the likes and dislikes of the researcher and,
possibly, data at hand. Several lines of research
can be recognized. However, two general
hypotheses emerge from the available literature.
The first hypothesis, which may be called
“push-pull,” is explained by structural changes
in U.S. agriculture. The second hypothesis tends
to explain part-time farming as a typical response
to industrialization and urbanization.
Most of the literature on part-time farming in
the U.S. was published in the 1930s, 1950s, and
1970s. During each wave of interest, similar
questions were asked and answered. Salter and
Diehl, in a survey article, characterized part-time
farming research in the 1930s as being “static and
descriptive” and stressed the problems arising
from lack of comparable definitions of a part-
time farmer. The article recommended more dy-
namic and analytical research.
The studies published in the 1950s and early
1960s can generally be classified into one of three
categories: (1) general descriptive-type studies
(Bauder; Fugitt; Galloway), (2) sociological stud-
ies (Fliegel; OECD), and (3) resource use or effi-
ciency studies (Jensen and Sundquist; Reinsel;
OECD). In the 1970s, several studies, such as
Bollman; Hanson and Spitze; Huffman; Singh
and Bagi, added to the knowledge and concept of
part-time farming and off-farm income. How-
ever, studies are needed to provide a better un-
derstanding of the incidence, characteristics, and
aspirations of part-time farmers in various re-
gions. There is also need for studies to determine
the extent to which a part-time farm’s production
costs and input-output coefficients differ from
those of a full-time farm and to investigate fur-
ther the implications (Carlin and Ghelfi; Bate-
man).
PRODUCTION EFFICIENCY DIFFERENCES
BETWEEN PART-TIME AND
FULL-TIME FARMS
Data and Methodology
Primary data used in this paper were obtained
in an enumerative survey of rural farm families in
two countries of western Tennessee in 1977-78.
The statistical analysis is based on data collected
from 193 randomly selected farm families repre-
senting 5.6 percent of all farm operators in the
two-county area.2 Out of a total of 193 farm
operators, 107 were classified as full-time, and 86
were classified as part-time.3 Personal interviews
were conducted with farm operators and data
were obtained on selected farm operations in the
previous year.
One method by which the economic efficiency
of farms can be analyzed is in the production
function framework. The economic efficiency
consists of two components—technical, and al-
locative or price efficiency. Overall economic ef-
ficiency, therefore, is a function of both price and
technical efficiency, and a firm is completely ef-
ficient economically only if it minimizes cost per
unit of output (Hall and LeVeen; Holland). Abso-
lute as well as relative allocative efficiency can
be analyzed in the production function frame-
work. However, technical efficiency is quite sen-
sitive to the specification of the production func-
tion. If one assumes, without testing, that the
underlying production function is linear homoge-
neous, he may be led to believe that the differ-
ences in allocative efficiency and in the configu-
ration of input and output prices are responsible
for any differences in yields and factor inten-
sities, while actually the answer lies in the tech-
nological differences among the distinct group of
farms (Barnum and Squire). Therefore, in this
study, we first examined the assumptions of
linearity and homogeneity of the production
function describing the nature of our sample
farms. The assumption OfHnearity is satisfied if
the elasticity of (returns-to-) scale is unity.
Hence, we estimated returns-to-scale, tested the
homogeneity assumption, and then proceeded to
analyze the technical and allocative efficiencies
of the selected farms.
In order to analyze the technical4 and alloca-
tive efficiencies on the selected farms, the follow-
ing log-linear Cobb-Douglas production function
was fitted:
(1) Ln Y = Ln C + In D + ɑj In L + α2 In N +
a3 In K + a4 In F + a5 In XL +
B1 (InL) *D + B2 (In N) *D +
B3 (In K) *D + B4 (In F) *D +
B5 (In XL) *D + u
where
Y = the value of crops, crop by-products,
2 For detailed data collection procedures and methodology, see Singh and Bagi.
3 For the purpose of this study, a part-time farm is defined as a “farm operated by an individual or partnership where the operator spends less than 50 percent of his
working time on the farm (does not consider farming to be principal occupation).’’ This was the definition used in the 1974 agriculture census to classify farms. At the time of
interview, the enumerator read the definition of a part-time farm and, if necessary, explained it to the operator. After the operator understood the definition, the answer was
noted, and the farm was classified as full- or part-time.
4 The economic efficiency has two components: technical and allocative efficiency. One of the anonymous reviewers correctly pointed out that, if all relevant inputs are
adequately measured, the technical efficiency coefficient will always be equal to one. But there may be errors of observation and measurement in output across farms. The
symmetrical random disturbance has been added to equation (1) to take care of the errors of observation of measurement on Y. However, the “technical efficiency”
coefficient also can be less than 1 ; and it may be variable across farms as a result of favorable as well as unfavorable external events, such as topography, soil type, machine
performance, luck, and the will and effort of the farmer (Aigner et al.). A one-sided, normal error term, in addition to u, can take care of such factors; but the estimation of
such a model would require complicated estimation methods. Therefore, our aim is a comparative analysis of the two farm groups, rather than an estimation of the “technical
efficiency” coefficients.
62