ECON321 Lecture Notes - Lecture 23: Probit, Heteroscedasticity, Autocorrelation
Document Summary
Usually we assume that causality is one way. Ex/ more education increases productivity which increases wages. Ex/ # of police per capita affects # of murder rates per capita. But number of murder rates per capita also affects the police force. Simultaneous causality bias- ols estimators will be biased and inconsistent in the presence of simultaneous causality. Y will be correlated with v (2) (1) 1 1 1 equation 3 is known as the reduced form equation. If y is correlated with v, then it will be correlated with . When y is correlated with v because of simultaneity, we say that the ols estimation of equation (1) suffers from simultaneity bias. Size and direction of the bias will depend on the covariance between y and v. When the sample selection is based on the dependent variable, then the error term and the independent variable(s) will be correlated, which leads to biased ols estimators.