The implications of estimating market demand curves by least squares regression
指出，即使统计上合理，用最小二乘法估计以数量为因变量的需求方程也会因归一化选择而低估需求弹性，并用美国肉类季度数据验证。
The use of least squares to estimate quantity-dependent demand equations has been justified on the grounds that the variance of the supply function dominates that of the demand function. However, the selection of least squares, even when statistically justified, introduces a normalisation decision, not generally discussed. It is shown that a least squares estimation of a quantity-dependent demand system underestimates demand elasticities as a consequence of the normalisation decision. This phenomenon is demonstrated for a quarterly model of the demand for beef, pork and poultry for the U.S.