Description

Abstract: Coffee is one of Vietnam’s most valuable exports, earning up to $600 million a year. However, a drop in producer coffee prices in the late 1990s has caused hardship for

4 million Vietnamese people who depend on the production of coffee for income. Also, Vietnam s coffee production has been associated with negative environmental externalities. Using data from 209 households in Dak Lak province, this thesis focuses on measuring the level of productive efficiency among coffee producers and identifying sources of inefficiency in the sample. Specific attention is focused on the importance of outliers in the dataset. A new, simple method for detecting outliers is introduced.

Initial findings suggest that almost half of the coffee farmers in the study sites are producing at an efficient level. In addition, Kinh farmers are less efficient than non-Kinh farmers. There are strong correlations between efficiency scores and farm size, labor, fertilizer, and pesticide. When interaction variables conditioned by ethnicity are added to the model, an additional positive and significant impact of education is found for ethnic minorities.

With regard to outliers, the thesis develops a new method to detect them based on the weights of observations in the dataset. After outliers are removed, levels of efficiency and sources of inefficiency are re-examined. Results from the new sample suggest that farm size is not significantly correlated with efficiency levels. All the other correlations remain unchanged. Economic inference regarding sources of inefficiency is found to be sensitive to the removal of outliers.