Endnotes

¹ Our analysis follows Battese and Coelli (1988, 1995); for further details see Coelli et al (1998).

²Significance herein refers to statistical significance.

³It should be noted that there is a high cash cost associated with organic fertilizers, which is interesting as we would normally assume that organic fertilizers are derived from farm wastes (such as compost) and therefore is appropriate for farmers who do not have access to credit. In the case of Thailand, it appears from the survey that there are commercial forms of organic fertilizers and farmers in the North do have cash to purchase these.

⁴An alternative is to include contract and non-contract farms in a single estimation and use a dummy variable to distinguish them. However, as pointed out by Delgado et al. (2003), Warnings and Key (2002), and Larsen and Foster (2005), such specifications may lead to self-selection or simultaneity bias since the decision to be a contract or organic farmer may not be independent from other production decisions.

⁵ Methodologically, our counterfactual simulations are based on a switching regression model (Maddala 1983, Chapter 8 and 9) and follow the two-stage estimation process suggested by Heckman (1976).
Let if farm i is a contract farm; and p_i = 0 otherwise. Then we first use the profit model to estimate a selection model specified as I^*_i = where I^*_i i i i i is a latent index capturing how farms choose between contract and non-contract farming; specifically farm i would choose contract farming (i.e. p_i = 1) if I^*_i > 0 non-contract farming (i.e. p_i = 0) if otherwise Z_i is farms’ characteristics that affect the probability of their choices between contract and non-contract farming.

⁶ Profit efficiency reported is an index adjusted by including nine negative profit observations that were dropped from the estimation The profit efficiency measure PE_iwhich measures the ratio of a farm’s actual profit to its maximum attainable profit, is not well defined when actual profits are negative. Since all the cases of negative profits are non-contract farms, excluding them would lead to biased results. Therefore, we apply the following measure of the profit efficiency of farms with negative actual profits. We first calculate the absolute value of profit loss of each of the nine negative profit farms compared to its estimated maximum attainable profit; let us denote such profit losses as Then, the profit efficiency of say farm i among these 9 negative-profited farms is measured by / max(), where max() represents the greatest profit loss among these 9 farms.

Under this profit efficiency measure, the profit efficiency score of a farm with negative profits would be negative and at the range of [-1, 0). The one with the largest profit loss would have profit efficiency score of -1; and the closer a farm’s negative profit efficiency to zero, the greater its profit efficiency score would be compared to other farms with negative actual profits. That farms with positive (or negative) actual profits have positive (or negative) profit efficiency scores implies that farms with negative actual profits must be less efficient than those with positive profits. This makes sense because farms with negative profits have lost more than whatever attainable profits they may have. Considering that we have used the least efficient farm as a benchmark to index the profit efficiency of farms with negative profits, we adjust the efficiency measure for positive profit farms accordingly by using PE_i / max(PE_j) to measure farm i’s efficiency. In sum, the adjusted profit efficiency scores are in the range of [-1, 1]. Farms with positive actual profits have positive profit efficiency scores, while farms with negative profits with negative scores. The greater a farm’s score is, the more profit efficient it is.

⁷ Similar to the estimation of counterfactual profits, we use the rice prices of non-contract farms to simulate contract farms’ counterfactual rice prices.

Download this Discussion Paper [ PDF 204.6KB| 29 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.