Skip to main content
. 2021 Jun;151:102626. doi: 10.1016/j.jdeveco.2021.102626

Table 6.

Treatment effects of each contract characteristic [T3-T2-T1-C].

OLS
OLS
ANCOVA
ANCOVA
(1) (2) (3) (4)
Panel A: rice area (ha)
Treatment effect of T1 0.247∗∗∗
(0.068)
0.228∗∗∗
(0.068)
0.247∗∗∗
(0.069)
0.228∗∗∗
(0.069)
Treatment effect of T2 0.134∗∗∗
(0.048)
0.115∗∗
(0.050)
0.133∗∗∗
(0.048)
0.114∗∗
(0.051)
Treatment effect of T3 0.268∗∗∗
(0.064)
0.251∗∗∗
(0.066)
0.268∗∗∗
(0.064)
0.251∗∗∗
(0.067)
Mean dependent variable in control 0.772
Observations 855 855 855 855
R-squared 0.079 0.085 0.080 0.085
Arrondissement FE Yes Yes Yes Yes
Household covariates No Yes No Yes

Panel B: yield (kg/ha)
Treatment effect of T1 520.4∗∗∗
(127.6)
518.8∗∗∗
(147.6)
508.9∗∗∗
(128.3)
506.6∗∗∗
(147.6)
Treatment effect of T2 424.9∗∗∗
(102.3)
453.5∗∗∗
(109.4)
418.8∗∗∗
(102.2)
447.6∗∗∗
(109.2)
Treatment effect of T3 502.8∗∗∗
(111.3)
506.5∗∗∗
(119.6)
494.5∗∗∗
(110.3)
497.8∗∗∗
(117.9)
Mean dependent variable in control 1652
Observations 855 855 855 855
R-squared 0.088 0.096 0.090 0.098
Arrondissement FE Yes Yes Yes Yes
Household covariates No Yes No Yes

Panel C: market participation (%)
Treatment effect of T1 22.11∗∗∗
(2.748)
23.87∗∗∗
(2.530)
22.12∗∗∗
(2.739)
23.89∗∗∗
(2.509)
Treatment effect of T2 31.43∗∗∗
(2.294)
32.28∗∗∗
(2.184)
31.43∗∗∗
(2.298)
32.29∗∗∗
(2.185)
Treatment effect of T3 39.05∗∗∗
(2.364)
40.76∗∗∗
(2.128)
39.05∗∗∗
(2.365)
40.77∗∗∗
(2.121)
Mean dependent variable in control 24.96
Observations 855 855 855 855
R-squared 0.530 0.542 0.530 0.542
Arrondissement FE Yes Yes Yes Yes
Household covariates No Yes No Yes

Panel D: income per capita (US$)
Treatment effect of T1 128.6∗
(69.98)
92.63
(72.76)
146.4∗∗
(69.61)
105.3
(71.87)
Treatment effect of T2 79.69
(73.42)
90.49
(74.70)
123.4
(75.62)
109.2
(74.83)
Treatment effect of T3 170.7∗∗
(68.78)
166.3∗∗
(68.63)
210.0∗∗∗
(67.24)
186.9∗∗∗
(67.51)
Mean dependent variable in control 265.3
Observations 855 855 855 855
R-squared 0.097 0.291 0.175 0.314
Arrondissement FE Yes Yes Yes Yes
Household covariates No Yes No Yes

Note: For simplicity, coefficient estimates are only reported for the treatment effect. Covariates include household size, age and gender of household head, number of years growing rice, and indicators for if the household head had at least primary education, if farming is the household’s main activity, if they have received extension training previously, and if they are a member of a farmer association. Standard errors clustered at the farmer-group-level are in parentheses (∗∗∗p<0.01, ∗∗p<0.05, ∗p<0.10).