Some of the questions we wish to investigate can be answered by careful compilation of data. Thus Chapter 6 directly addressed hypotheses 1 and 6 of Chapter 1; these concerned (respectively) which size classes of producers have higher profits per unit of output, and what the effect of contract farming is on this finding. Furthermore, hypothesis 3, concerning which size class of farmers make more of an effort to mitigate negative environmental externalities per unit of output, was dealt with in Chapter 7. Answering the other hypotheses introduced in Chapter 1 requires econometric estimation, as laid out in Chapter 4.
This is particularly true of issues involving efficiency, which require establishing an ideal benchmark (a frontier) customized to each farm's resource endowment and price environment, in order to place each farm's actual profit performance in the context of the ideal applicable to that farm, in a way suitable for comparison across farms of different sizes. It is also the case for any questions that require assessing which factors determine (explain) outcomes such as farm-specific environmental mitigation behavior. It is even more necessary to have a clear econometric approach when trying to sort out myriad influences that are simultaneously affecting a composite result such as a single indicator of farm-specific profit efficiency.
The present chapter synthesizes the key results of econometric models that were in most cases estimated separately for each commodity in each country. This involves thirteen separate models: India broilers, layers and dairy; Philippines swine and broilers; and broilers, layers, dairy and swine in both Brazil and Thailand. Then separate runs in some cases were done for large and small farms within commodity/country categories, and for contract farmers. Separation of estimation by country and commodity was practiced because pooling across commodities or countries in our case would either introduce biases in the estimation that would be very difficult to control for, or remedies such as fixed effects and interaction terms that would obliterate in many cases the fine-tuning necessary to get insights on the most relevant issues.
Cross-country analysis is organized by commodity group. The synthesis is presented in terms the maximum aggregation that makes sense. For example, on issues involving environmental mitigation, broilers and layers are close enough to be combined, but need to be separated fro analysis of efficiency. Widespread contract farming was observed only for some commodities in some countries, and additional econometric insights from including contract farmers are dealt with in separate sections.