Conley, Hansen, and Rossi proposed a "plausible" form of this property in which a parameter symbolized gamma is close to zero, but not necessarily exactly zero. In the order condition of simultaneous equation models, the number of excluded variables with this property must be at least the number of included variables lacking this property. A deviation from this property originating from model specification errors can be found with the Durbin-Wu-Hausman test. An instrumental variable must have this property to satisfy the (*) exclusion restriction condition. A common source of deviation from this property is omitted variable bias if the omitted variable is correlated with variables of interest. (-5[1])If the variables have this property in a regression (-5[1])model, the model is unbiased. (-5[3])For 10 points, (-5[1])name (-5[2])this property in which a variable is uncorrelated (-5[2])with the error, (-5[5])which is contrasted with endogeneity. (10[3]) (10[12]0[2])■END■