Effect of adoption of organic farming on technical efficiency of olive-growing farms: empirical evidence from West Bank of Palestine

Agricultural and Food Economics

Table 9 Average treatment effects: Endogenous switching regression model using instruments with heteroscedasticity.

Outcome variables	Farm type and treatment effect	Decision stage
Outcome variables	Farm type and treatment effect	Adopt	Not to adopt	Average treatment effect	Standard error
Metaefficiency	Organic farms (ATT)	0.698	0.579	0.119^***	0.017
Metaefficiency	Conventional farms (ATU)	0.690	0.599	0.091^***	0.013

We estimated the ESR using the instruments with heteroscedasticity proposed by Lewbel (2012, 2018a, 2018b). Following empirical application by Lin et al. (2022), we run the regression of the endogenous variable (adoption of organic farming) on the exogenous variable Z_i and predict error term e_i. Then, instrumental variables are constructed as \(\left({{\varvec{Z}}}_{i}-\overline{{\varvec{Z}} }\right){e}_{i}\), where \(\overline{{\varvec{Z}} }\) is the sample mean of Z_i. The elements of Z_i are all control variables from Eq. (8). This method assumes that residual e_i is supposed to be heteroscedastic. Standard errors were obtained by bootstrapping with 300 replications.
*, **, *** indicate significant at the 10%, 5%, 1% level, respectively