Skip to main content

Table 9 Multinomial endogenous treatment effects model estimates

From: IITA’s genebank, cowpea diversity on farms, and farmers’ welfare in Nigeria

Explanatory variables Outcome models
Ln cowpea yield Ln cowpea consumption Ln cowpea sale
Treatment variable
 Adoption of cowpea variety without genebank ancestry − 1.545** (0.781) 0.186 (0.205) − 0.578 (1.025)
 Adoption of cowpea variety with genebank ancestry 1.019*** (0.395) 0.381*** (0.095) − 0.027 (0.158)
Household characteristics
 Age − 0.018 (0.022) 0.012 (0.010) 0.001 (0.035)
 Age2 0.000 (0.000) − 0.000 (0.000) − 0.000 (0.000)
 Sex 0.735*** (0.238) 0.081 (0.101) − 0.116 (0.258)
 Education 0.007 (0.025) − 0.013 (0.010) 0.017 (0.041)
 Education2 0.000 (0.002) 0.000 (0.001) − 0.002 (0.003)
 Need credit 0.385*** (0.109) 0.059 (0.045) − 0.631*** (0.132)
Farm characteristics
 Size − 0.172*** (0.031) 0.065*** (0.015) − 0.000 (0.021)
 Slope 0.041 (0.056) − 0.044* (0.025) 0.105 (0.071)
 Distance to farm − 0.006*** (0.002) 0.001 (0.001) 0.003 (0.002)
 Soil fertility − 0.335** (0.151) − 0.134** (0.067) 0.133 (0.220)
Market characteristics
 Distance to village market − 0.003*** (0.001) − 0.000 (0.000) − 0.004*** (0.001)
 Distance to district market − 0.001* (0.001) 0.000 (0.000) 0.000 (0.001)
 Constant 3.111*** (0.620) 0.822*** (9.264) 5.399*** (0.909)
Selection terms \((\lambda )\)
 Adoption of cowpea variety without genebank ancestry 0.958** (0.480) − 0.419*** (0.156) 0.000 (0.000)
 Adoption of cowpea variety with genebank ancestry − 1.194*** (0.457) − 0.419*** (0.109) − 0.000 (0.000)
 Prob > chi2 0.000 0.000 0.000
 N 1442 1376 500
  1. Columns 1, 2 and 3 are multinomial endogenous treatment effects model regressions with natural logarithms of cowpea yield, cowpea consumption and cowpea sale as outcomes, respectively. The control group is households that grow a cowpea landrace. 1000 simulation draws were used. Standard errors are in parentheses. ***,**,*: Significance at 1, 5, or 10% significance level, respectively