namely, production risk, price risk and input risk. Production risk may depend on
the region in which the farmer produces crop. This is because rainfall as well as soil
type varies across regions. To see whether region or production risk plays an
important role in determining crop insurance, we have considered a dummy variable, where we have assigned a value 1 to households belonging to inland North
region of Karnataka, zero value has been assigned otherwise. Inland north region of
Karnataka is a dry semi-arid region and it includes northern districts of Karnataka,
such as Bagalkot, Bijapur, Gulbarga, Bidar, Dharwad, Haveri, Chitradurga,
Davangere, Gadak, Raichur, Koppal, etc.
Apart from production risk, a household may face price risk and input risk. Since
price risk and input risk varies across crops, we have included dummy variables to
capture it. Generally, cereal crops are expected to have less fluctuation in price, so
we have assigned a value 1 if the farmer produces mainly cereal crop, zero value
has been assigned otherwise. However, a significant relation between type of crop
and crop insurance does not necessarily imply inducement of price risk to avail
insurance. The insurance terms and conditions may be better for some crops, which
in turn may induce households in having insurance for that crop.
In addition to the above-mentioned variables, risk faced by a household is
expected to be less if a household has diversified into non-farm activities.
Households which have diversified into non-farm activities are captured in the
explanatory part with the help of a dummy variable.
Awareness about crop insurance may depend on the education of the household.
Educated household are expected to have better information. Therefore, we have
assigned a value 1 to households having at least one member with secondary
education, zero value is assigned otherwise. Age can also play an important role in
creating awareness. Households with more aged people are expected to have more
awareness. In the present analysis, mean age of the household is captured as an
explanatory variable.
Information on insurance benefits in India may vary with caste and
gender (Rajeev et al. 2011). Households with female head and households
belonging to poorer caste may have poorer social networking with respect to
insurance leading to less presence of information.
Supply side factors are expected to vary across regions, which is already captured. For example, in semi-arid region, supply of insurance may be less because
higher risk of crop failure.
3.2 Probit Regression
In Probit model it is assumed that the zero or one value assumed by the observed
dependent variable, depends an unobserved latent variable ðY
Ã
i Þ: such that if Y
Ã
i
exceeds a critical value (I i ), a household avails insurance and we observe a value 1
and if Y
Ã
i is less than this threshold value, a household is without crop insurance.
Climate Change and Uncertainty in Agriculture …
247
the region in which the farmer produces crop. This is because rainfall as well as soil
type varies across regions. To see whether region or production risk plays an
important role in determining crop insurance, we have considered a dummy variable, where we have assigned a value 1 to households belonging to inland North
region of Karnataka, zero value has been assigned otherwise. Inland north region of
Karnataka is a dry semi-arid region and it includes northern districts of Karnataka,
such as Bagalkot, Bijapur, Gulbarga, Bidar, Dharwad, Haveri, Chitradurga,
Davangere, Gadak, Raichur, Koppal, etc.
Apart from production risk, a household may face price risk and input risk. Since
price risk and input risk varies across crops, we have included dummy variables to
capture it. Generally, cereal crops are expected to have less fluctuation in price, so
we have assigned a value 1 if the farmer produces mainly cereal crop, zero value
has been assigned otherwise. However, a significant relation between type of crop
and crop insurance does not necessarily imply inducement of price risk to avail
insurance. The insurance terms and conditions may be better for some crops, which
in turn may induce households in having insurance for that crop.
In addition to the above-mentioned variables, risk faced by a household is
expected to be less if a household has diversified into non-farm activities.
Households which have diversified into non-farm activities are captured in the
explanatory part with the help of a dummy variable.
Awareness about crop insurance may depend on the education of the household.
Educated household are expected to have better information. Therefore, we have
assigned a value 1 to households having at least one member with secondary
education, zero value is assigned otherwise. Age can also play an important role in
creating awareness. Households with more aged people are expected to have more
awareness. In the present analysis, mean age of the household is captured as an
explanatory variable.
Information on insurance benefits in India may vary with caste and
gender (Rajeev et al. 2011). Households with female head and households
belonging to poorer caste may have poorer social networking with respect to
insurance leading to less presence of information.
Supply side factors are expected to vary across regions, which is already captured. For example, in semi-arid region, supply of insurance may be less because
higher risk of crop failure.
3.2 Probit Regression
In Probit model it is assumed that the zero or one value assumed by the observed
dependent variable, depends an unobserved latent variable ðY
Ã
i Þ: such that if Y
Ã
i
exceeds a critical value (I i ), a household avails insurance and we observe a value 1
and if Y
Ã
i is less than this threshold value, a household is without crop insurance.
Climate Change and Uncertainty in Agriculture …
247