Impact of Weather Shock on Food Insecurity: A Study on India
343
5 Regression Results
The results of the binary logistic regression of food insecurity are shown in Table 4.
Here the odds ratios are reported rather than the coefficients, and the results are
interpreted accordingly.
2 It may be noted that the coefficients of all the explanatory
variables except other caste (OC) is statistically significant which implies that they
have statistically significant impacts on our dependent variable.
The odds ratio of weather shock (WS) has turned out to be more than 1 which
signifies a positive impact of weather shock on the probability of food insecurity. This
implies that if deviations from normal rainfall in an area increase, then the households
belonging to that area tend to be more food insecure. This is quite obvious because
weather shocks, either in the form of excessive rainfall or deficient rainfall, are
harmful for the agriculture sector which reduces output.
Table 4 Results of the binary
logistic regression
Explanatory variables
Odds ratio
Std. error
Weather shock (WS)
1.003***
0.000
Per capita income (PCI)
0.999***
0.001
Household size (HS)
1.203***
0.008
Dependency ratio (DR)
1.158***
0.022
Remittances (Remit)
0.921*
0.042
Poor (Pov)
12.512***
0.481
Main occupation (Cultivation)
0.784***
0.028
Female adult education (FAE)
0.989***
0.003
Residence (Ur)
1.076**
0.039
Religion (Hindu)
1.200***
0.047
Other backward classes (OBC)
1.494***
0.058
Scheduled castes (SC)
1.634***
0.073
Scheduled tribes (ST )
2.091***
0.119
Other castes (OC)
1.040
0.146
Const
0.097***
0.006
Pseudo R squared
28.94%
LR chi 2 (15)
12326.82
Prob > chi 2
0.0000***
Observations
35301
Note ***, ** and * represent significant at 1%, 5% and 10%
respectively
2 This is because there is a direct relationship between the two. Odds ratios greater than 1 and less
than 1 imply positive and negative coefficients respectively. Therefore, our analysis is carried out
in terms of odds ratio as it is easier to interpret the impact of the explanatory variables in terms of
odds ratio.
343
5 Regression Results
The results of the binary logistic regression of food insecurity are shown in Table 4.
Here the odds ratios are reported rather than the coefficients, and the results are
interpreted accordingly.
2 It may be noted that the coefficients of all the explanatory
variables except other caste (OC) is statistically significant which implies that they
have statistically significant impacts on our dependent variable.
The odds ratio of weather shock (WS) has turned out to be more than 1 which
signifies a positive impact of weather shock on the probability of food insecurity. This
implies that if deviations from normal rainfall in an area increase, then the households
belonging to that area tend to be more food insecure. This is quite obvious because
weather shocks, either in the form of excessive rainfall or deficient rainfall, are
harmful for the agriculture sector which reduces output.
Table 4 Results of the binary
logistic regression
Explanatory variables
Odds ratio
Std. error
Weather shock (WS)
1.003***
0.000
Per capita income (PCI)
0.999***
0.001
Household size (HS)
1.203***
0.008
Dependency ratio (DR)
1.158***
0.022
Remittances (Remit)
0.921*
0.042
Poor (Pov)
12.512***
0.481
Main occupation (Cultivation)
0.784***
0.028
Female adult education (FAE)
0.989***
0.003
Residence (Ur)
1.076**
0.039
Religion (Hindu)
1.200***
0.047
Other backward classes (OBC)
1.494***
0.058
Scheduled castes (SC)
1.634***
0.073
Scheduled tribes (ST )
2.091***
0.119
Other castes (OC)
1.040
0.146
Const
0.097***
0.006
Pseudo R squared
28.94%
LR chi 2 (15)
12326.82
Prob > chi 2
0.0000***
Observations
35301
Note ***, ** and * represent significant at 1%, 5% and 10%
respectively
2 This is because there is a direct relationship between the two. Odds ratios greater than 1 and less
than 1 imply positive and negative coefficients respectively. Therefore, our analysis is carried out
in terms of odds ratio as it is easier to interpret the impact of the explanatory variables in terms of
odds ratio.