Glossary of economics research

Results of search for identification follow:

identification: A parameter in a model is identified if and only if complete knowledge of the joint distribution of the observed variables would be enough information to calculate the parameter exactly.

If the model has been written in such a way that its parameters can be consistently estimated from the observables, then the parameters are identified. There exist cases (mostly obscure) where parameters are identified but consistent estimators are not possible, as shown in this discussion drawn from the elegant paper of Gabrielsen (1978).

A model is identified if there is no observationally equivalent model. That is, potentially observable random variables in the model have different distributions for different values of the parameter.

Formally:
Let h* be a vector of unknown functions and distributions in an econometric model.
Let H denote a set which h* is known to belong. H is defined by the model's restrictions.
Let P(h) denote the joint distribution of observable variables of the model for various elements of h in H. The distribution for the actual data will be assumed to be P(h*).
Now, vector h* is identified within H if for all h in H such that h<>h* it is true that P(h)<>P(h*).

Note: Linear models are either globally identified or there are an infinite number of observably equivalent ones. But for models that are nonlinear in parameters, "we can only talk about local properties." Thus the idea of locally identified models, which can be distinguished in data from any other 'close by' model.

"An identification problem occurs when a specifed set of assumptions combined with unlimited observations drawn by a specified sampling process does not reveal a distribution of interest." -- Manski, Charles F. "Identification problems and decisions under ambiguity: empirical analysis of treatment response and normative analysis of treatment choice" Northwestern University Department of Economics and Institute for Policy Research, September 1998, p. 2

Source: The New Palgrave: Econometrics, p 96; and Gabrielsen, 1978
Contexts: econometrics


Back to top

Type 'help' for a list of relevant categories. Use * as a general wild card or ? for one character.

Feedback to econterms at econterms.com