Expectation Formation and Financial Ratio Adjustment Processes: A Comment and An Extension.
指出Lee和Wu（1988）使用高斯-牛顿非线性回归估计财务比率调整模型的方法存在参数模糊问题，并提出基于最大似然的分布滞后估计方法，结果与原文不一致。
Abstract Lev (1969) uses a partial adjustment model to examine whether firms adjust their financial ratios to industry averages. Lee and Wu (1988) extend Lev's model by introducing adaptive expectations into the partial adjustment model to examine whether firms adjust their desired target financial ratios. Lee and Wu examine six ratios of 112 U.S. firms from eight industries for the period 1965-1984. Their primary results show that adjustments to target ratios are not instantaneous. In this comment, we show that Lee and Wu's (1988) use of the Gauss-Newton nonlinear regression method to estimate the structural parameters of the autoregressive form of their partial adjustment-adaptive expectations model is inappropriate. The problem is ambiguity in their estimates, because the two coefficients (the speed of adjustment and the reaction of expectations) in the model occur symmetrically (Maddala 1988, 353). Consequently, Lee and Wu's parameter estimates are sensitive to the selection of the initial parameter values used in their estimation. Furthermore, for the six financial ratios Lee and Wu examine, about 10 to 25 percent of their estimates of the coefficients for the speed of adjustment and approximately 20 to 30 percent of their estimates of the coefficients for the reaction of expectations exceed their theoretical upper limit of one. To overcome these problems, we propose an alternative method by estimating the distributed lag form of the model using a maximum-likelihood-based approach (Maddala 1988, 343-44; Kmenta 1986, 532-33). Using this procedure, we obtain results that are inconsistent with those of Lee and Wu (1988).