4,249 works (including 483 books, 3,406 articles and 357 historical documents), 150 years of economic thought. Each one summarized and searchable.
Even complete knowledge of a system’s observable behavior may leave its underlying parameters undetermined. In this survey contribution, Manfred Deistler and Gerhard Tintner examine that obstacle before estimation begins, linking measurement-error models with multivariate ARMA systems. Their perspective brings probability theory into contact with the algebra of dynamic models: nonnormal distributions and lagged covariances can supply identifying information, while different polynomial representations can describe exactly the same observable process. Crucially, they distinguish restrictions expressing substantive knowledge from canonical conventions that merely select a unique representation. Readers can discover why more data alone cannot resolve some ambiguities—and why even a uniquely specified model needs a well-behaved parameterization if identification is to support consistent estimation.