Table 3.

Description of the sampled households and respondents.

Variable	Male (n = 299)	Female (n = 134)	Diff.	Z or χ2 value	P
Household characteristics
Religion (non-Muslim) (%)	1%	7%	6%	9.19	<0.01
Household size	6.58	5.57	1	4.13	<0.001
Cultivable land (decimal)	103	45	58	4.67	<0.001
Non-land asset (in US$)	1973	1245	728	3.28	<0.001
Average size of maize farm in decimal (hectare)	31 (0.14)	26 (0.11)	4.72	1.17	0.30
Cost (median) of production of per hectare of maize farm in 2014 (in USD)	476	487	11	20478a	0.71
Revenue (median) earned from per hectare of maize farm in 2014 (in USD)	1191	1016	175	16189a	<0.001
Profit (median) earned from per hectare of maize farm in 2014 (in USD)	667	476	190	15925a	<0.001
Maize cultivation experience (years)	3.50	3.0	0.47	2.50	<0.05
Formal savings account (%)	43	49	−6	1.40	0.30
Formal credit account (%)	42	54	−12	4.72	<0.05
Purchased insurance (%)	21	22	−1	0.018	0.88
Respondent characteristics
Mean age (years)	45	35	9.5	7.40	<0.001
High school and above (%)	35	22	13	7.33	<0.01
Head of the household (%)	87	23	64	171	<0.001
No familiarity with insurance (%)	53	63	−10	3.70	<0.10
Risk aversion coefficientb	0.73	0.91	−0.23	2.45	<0.05
Time preferencec (% with discount rate >70%)	86	68	18	21	<0.001

^aMann-Whitney U statistics.

^bAssuming constant relative risk aversion (CRRA), $u\left(y\right)=y^{1-\theta }/1-\theta$, the curvature of the utility function θ represents the degree of risk aversion. This was determined by calculating the value of θ that would make a respondent indifferent between the chosen gamble and the two adjacent gambles (Eckel and Grossman, 2008). The mean of the risk-aversion coefficient is 0.78, which is consistent with CRRA risk-aversion coefficients for farmers in developing countries (Olbrich et al., 2012).

^cThe discount rate is determined solving the value function $v\left(M_0\right)=\frac{1}{1+r}v\left(M_t\right)$. M₀ is the present value of M_t offered at time t with discount rate r.