Applying Theil's Multinomial Extension of the Linear Logit Model to Meat Expenditure Data
针对多项线性Logit模型线性化后参数缺乏经济含义的问题，提出替代变换方法使系数可解释为弹性或斯拉茨基参数，并用于英国肉类需求实证分析。
Abstract One of the chief drawbacks of the multinomial extension of the linear logit model is that, after the standard linearization, the parameters which are directly estimated have no obvious economic meaning. This paper presents alternative transformations which facilitate interpretation. In one linearization the estimated coefficients will be elasticities at a selected data point such as the mean. In another form, the coefficients will be Slutsky parameters. The latter version is chosen for an empirical analysis of the demand for meat in Great Britain. A number of issues concerning the model's error structure and dynamic specification are discussed.