Panel Data Estimation Techniques and Farm‐level Data Models
说明在农业政策分析中，使用更复杂的面板数据技术（如处理异方差和误差项相关性的两阶段方法）能提高参数估计的可靠性，对使用农场层面数据的政策研究者有参考价值。
Econometric models estimating parameters for agricultural policy analysis increasingly rely on unbalanced panels of farm‐level data. Since such models have often been estimated through simplified approaches, in this paper we show that adopting more sophisticated panel data techniques may be very important for obtaining more reliable estimates of policy parameters. We also extend the two‐stage procedure proposed by Shonkwiler and Yen (1999) and Tauchmann (2005) for the analysis of censored data to account for heteroskedasticity and correlation of the error terms of the first‐stage probit models.