Joint determination of the choice of planting season and technical efficiency of potato in northern Thailand: A comparison of Greene’s versus Heckman’s sample selection approach

Abstract

The potato sector in Thailand has undergone significant changes with rapid expansion of the contract farming system, thereby, leading farmers to grow potato either early or late in the season. The present paper addresses two interrelated issues: (a) jointly determine the choice of planting season, productivity and technical efficiency of potato cultivation in northern Thailand; and (b) compare suitability of the two competing methods to analyze the issues, that is, Heckman’s switching regression versus Greene’s sample selection in stochastic frontiers. Results obtained are different but Greene’s model fare better in identifying selectivity bias and model fit. The decision to plant potato early in the season is significantly determined by experience gained through contract farming and expected price, consistent with expectation. Land area and seed are the significant determinants of potato productivity and foliage hormone reduces productivity. The mean technical efficiency (MTE) in Greene’s model is lower by 5 points (p<0.01) than Heckman’s (that is, 0.80 versus 0.85). The main difference is at the upper end of the efficiency distribution. Policy recommendation includes keeping potato price high, promotion of contract farming system and distribution of potato seeds which will synergistically increase adoption and productivity of early season potato and raise farmers’ income. Key words: Technical efficiency, stochastic frontier, sample selection, potato, Thailand, contract farming, switching regression.