In statistics and econometrics, extremum estimators are a wide class of estimators for parametric models that are calculated through maximization (or minimization) of a certain objective function, which depends on the data. The general theory of extremum estimators was developed by .
An estimator is called an extremum estimator, if there is an objective function such that
where ÃÂ is the parameter space. Sometimes a slightly weaker definition is given:
where o<sub>p</sub>(1) is the variable converging in probability to zero. With this modification doesn't have to be the exact maximizer of the objective function, just be sufficiently close to it.
The theory of extremum estimators does not specify what the objective function should be. There are various types of objective functions suitable for different models, and this framework allows us to analyse the theoretical properties of such estimators from a unified perspective. The theory only specifies the properties that the objective function has to possess, and so selecting a particular objective function only requires verifying that those properties are satisfied.
If the parameter space àis compact and there is a limiting function Q<sub>0</sub>(ø) such that: converges to Q<sub>0</sub>(ø) in probability uniformly over ÃÂ, and the function Q<sub>0</sub>(ø) is continuous and has a unique maximum at ø = ø<sub>0</sub> then is consistent for ø<sub>0</sub>.
The uniform convergence in probability of means that
The requirement for àto be compact can be replaced with a weaker assumption that the maximum of Q<sub>0</sub> was well-separated, that is there should not exist any points ø that are distant from ø<sub>0</sub> but such that Q<sub>0</sub>(ø) were close to Q<sub>0</sub>(ø<sub>0</sub>). Formally, it means that for any sequence {ø<sub>i</sub>} such that , it should be true that .
Assuming that consistency has been established and the derivatives of the sample satisfy some other conditions, the extremum estimator converges to an asymptotically Normal distribution.