Agriculture Reference
In-Depth Information
to adopt drip irrigation. For this purpose, area under drip irrigation installed by the
farm households was considered as the dependent variable. It is expected that the
adoption of drip irrigation by the farm households influenced by different physical,
socioeconomic, institutional and household specific factors.
The dependent variable adoption of drip irrigation would be zero for those house-
holds who do not adopt drip irrigation. If the dependent variable is censored, values in
a certain range may all be recorded as single value. Given that our dependent variable
is censored at zero, a Tobit estimation rather than OLS is appropriate [7, 21]. In such a
case, Tobit estimators may be used. Thus, the functional form of the model specifi ed in
the present study with a Tobit model, with an error term (Ui) which is independently,
normally distributed with zero mean and constant covariance, is given by Eq. (1).
DA
*
i = Xi b + Ui For i = 1.n
DAi = T
*
i if Xi b + Ui > 0
= 0 if Xi b + Ui <= 0
(1)
In Eq. (1): DAi = Area under drip irrigation in hectares; Xi = Vector of independent
variables; b = Vector of unknown coefficients; and n = Number of observations. In the
functional relationship defined by Eq. (1), the DAi is the endogenous variable which
is expected to influence by other exogenous variables viz., age of the farmer in years
(AGE), educational level of the farmer in years of schooling (EDUCATION), farm
size in hectares (FSIZE), proportion of wider spaced crop (WIDERCROP), participa-
tion in nonfarm income activities (NONFARM) and percentage of area irrigated by
wells (AWELLS).
Economic implications can be drawn by using the results of the empirical model.
Following a Tobit decomposition framework suggested by McDonald and Moffi tt
[10], the effects of the changes in the explanatory variables on the probability of adop-
tion of drip irrigation and intensity of adoption could be obtained.
The basic relationship between the expected value of all observations, E(DA), the
expected value conditional upon being above the limit, E(DA
*
), and the probability of
being above the limit, F(
z
), is defi ned by Eq. (2).
(2)
*
EDA
(
)
=
EDA Fz
(
).
( )
The effect of a given change in the level of the explanatory variables on the dependent
variables (the Eq. (3)) can be obtained by decomposing the Eq. (2).
∂
EDA
(
)
⎛
∂
EDA
(
*
)
⎞
⎛
∂
Fz
( )
⎞
*
=
Fz
()
+
EDA
(
)
(3)
⎜
⎟
⎜
⎟
∂
X
⎝
∂
X
⎠
⎝
∂
X
⎠
i
i
i
Thus, the total elasticity of change in the level of the explanatory variable consists of
two effects: (i) change in DA of those above the limit (i.e., elasticity of intensity of
drip adoption, for those households who already an adopter) and (ii) the change in the
probability of being above the limit (i.e., probability of drip adoption).
Search WWH ::
Custom Search