Obstacles to the profitable production and marketing of horticulture products in Lesotho: an offset-constrained probit modelling of farmers’ perception - Institutional Constraints to Small Farmer Development in Southern Africa - page 162

of title deeds to the land). Importantly, only about a quarter of the surveyed households
expressed satisfaction with the results they obtain from their farming activities in respect of
horticultural production. There was also evidence that the households experienced serious
problems with marketing of produce where nearly half the sample had serious transport
problems and farmers were about 4 kilometres away from the nearest market.

Due to suspected multi-colinearity, the software dropped all but the 7 variables presented
in Table 8.3. The variable for membership of farmer associations was also dropped, making
it impossible to assess the influence of this variable which is intuitively considered useful
and will be investigated separately. The probit model to determine the impact of the set of
institutional factors on the attitude of the farmers to the profitability of the horticulture
sector is presented in Table 8.3. The intention was to show the probability that the farmer
would be satisfied with the production performance of his/her horticulture farming as a
result of a number of institutional influences in the farming environment. The analysis
imposed revenue per hectare as an offset which could still be influential despite the
difficulty in directly modelling that variable in a society where there are valid concerns over
its reliability when based on farmer's recall. The results are presented in Table 8.3.

According to the results, the probit model was highly significant and suggested very good
fit, with a Wald Χ 2 of 514.42 and log likelihood of -1,119.79 ( P =0.000). This would mean
that the modelled variables, with the exception of standards/grading, had strong enough
influence to determine whether or not the farmers would be satisfied with the way the
outcomes of their farming enterprises. The marginal effects are displayed in Table 8.4 and
suggest that there was probably a 9% higher probability of a farmer being satisfied with
horticulture production if s/he had land access than if access was non-existent. In general,

Table 8.3. Log-likelihood estimates and goodness-of-fit measures for the identified market and
institutional factors.

Variable Coefficient Standard error z-value

P -value

Standard/grading -0.118 0.315 -0.38 0.707

Land access 0.822 0.295 2.79 0.005

Transport facilities 3.401 0.407 8.36 0.000

Title deeds 4.840 0.304 15.93 0.000

Market distance 0.175 0.087 2.01 0.045

Transport problems -1.182 0.323 -3.66 0.000

Extension visit 3.102 0.509 6.09 0.000

Number of observations: n = 100; Wald chi 2 (7) = 514.42; Log likelihood = -1,119.7855; Prob > chi 2 = 0.0000.