Table 7. Joint effect of climate and health shocks on coping strategies against expenditure following health shock, Sundarbans, West Bengal, India.

Coping against medical and non-medical expenditures following health shock#	Health Shocks alone (n = 45)	Multiple Shocks (n = 49)	Total Sample @	Diff. in Mean	t-statistic	Odds Ratios on Multiple Shock parameter∧	Pseudo R2
(A)	(B)	(C)	(D)	(E)	(F)	(G)	(H)
Difficulty in financing health expenses	53.3	83.7	69.1	−29.6***	−3.332	3.837** (2.471)	0.331
Financing through household income and/or dis-saving	68.9	48.9	58.5	20.0**	1.9769	0.263** (0.149)	0.196
Financing through informal debts and credit	35.6	59.2	47.9	−23.6**	−2.3324	4.623* (2.892)	0.344

Note: ^#Expenditure items include direct costs of treatment, expenses on drugs and medicines, transport, and related expense such as those incurred on food or lodging for patients and accompanying person(s).

@ denotes that the testing of hypotheses for difference of means (results in column B-F) was carried out only for the households reporting an incidence of health shock (n = 94).

^∧The coefficients in column G are odds-ratios on the ‘multiple-shock’ variable – households experiencing both the health shock as well as high-impact due to climatic shock – from logit regressions with the dependent variable in the corresponding row of the first column. Health shock variable is for illness of household head and/or spouse. The comparison group for column G coefficients is households with health shock alone (with a less-impact of Climatic shock). Models additionally control for (log) total health expenses, pre-shock vulnerabilities (see text) and village-level fixed effects. Coping models (items B and C, in column A) also controls for the self-assessed ‘difficulty in financing’ variable.

Figures in parentheses are t-statistics testing for the hypothesis that the variable is not different from zero.

* *p<0.05, *** p<0.01.