The Estimation of a Simultaneous Equation Generalized Probit Model

Earnings Guidance and Managerial Myopia

“…If the instruments themselves are endogenous, it leads to bias in the twostage coefficients. In case of over-identified models such as ours (i.e., where the number of instruments exceeds the number of endogenous regressors), it is possible to regress the estimated 14 Amemiya [1978] shows that although the coefficients from the Nelson-Olsen second-stage equation are asymptotically unbiased, they are not the most efficient estimators and the covariance matrix of the estimators are biased. He proposes a method for deriving more efficient estimators and a correction to the covariance matrix that is unbiased.…”

Change of Scale and Forecasting with the Control-Function Method in Logit Models

“…That is, the ratios between the estimators are consistent estimators of the ratios of the parameters of the true model, but the actual estimators of the model parameters are inconsistent. This is also true, in general, for other methods to correct for endogeneity in discrete choice models, such as BLP (Berry et al, 1995), 2SIV (Newey, 1985 and Amemiya's (1978) methods.…”

“…Other methods that have been proposed are 2SIV (Newey, 1985) and Amemiya's (1978) method. The control-function method is particularly suitable to address endogeneity in several models that are relevant in transportation research.…”

National exchange rate policies and international debt crises: how Brazil did not follow Argentina into a default in 2001-2002

“…This alternate relationship has already been posited by Blanchard (2004), and Dreher et al (2005), and Herz and Tong (2003) have also considered this reverse causality in greater depth. One possibility would be to re-estimate the model according to the two-equation simultaneous equation model developed by Amemiya (1978). He proposes an estimation using two dependent variables in which one is observed in full, while the other is only observed to the extent that it is positive.…”

“…While a currency crisis variable is not used, it would to take values of 0 and 1 only like the debt crisis variable, so that the results could indicate the probabilities of either debt or currency crises occurring. As a result, these two dependent variables would not fall under the rubric set out by Amemiya (1978), and the alternate results could be deceptive.…”