Next Article in Journal
Consistency in Estimation and Model Selection of Dynamic Panel Data Models with Fixed Effects
Previous Article in Journal
Bayesian Approach to Disentangling Technical and Environmental Productivity
Open AccessArticle

A New Approach to Model Verification, Falsification and Selection

Department of Economics Temple University, Philadelphia, PA 19122, USA
*
Author to whom correspondence should be addressed.
Academic Editor: Kerry Patterson
Econometrics 2015, 3(3), 466-493; https://doi.org/10.3390/econometrics3030466
Received: 10 March 2015 / Revised: 7 June 2015 / Accepted: 16 June 2015 / Published: 29 June 2015
This paper shows that a qualitative analysis, i.e., an assessment of the consistency of a hypothesized sign pattern for structural arrays with the sign pattern of the estimated reduced form, can always provide decisive insight into a model’s validity both in general and compared to other models. Qualitative analysis can show that it is impossible for some models to have generated the data used to estimate the reduced form, even though standard specification tests might show the model to be adequate. A partially specified structural hypothesis can be falsified by estimating as few as one reduced form equation. Zero restrictions in the structure can themselves be falsified. It is further shown how the information content of the hypothesized structural sign patterns can be measured using a commonly applied concept of statistical entropy. The lower the hypothesized structural sign pattern’s entropy, the more a priori information it proposes about the sign pattern of the estimated reduced form. As an hypothesized structural sign pattern has a lower entropy, it is more subject to type 1 error and less subject to type 2 error. Three cases illustrate the approach taken here. View Full-Text
Keywords: verification; falsification; specification; qualitative analysis; entropy; inherited restrictions verification; falsification; specification; qualitative analysis; entropy; inherited restrictions
MDPI and ACS Style

Buck, A.J.; Lady, G.M. A New Approach to Model Verification, Falsification and Selection. Econometrics 2015, 3, 466-493.

Show more citation formats Show less citations formats

Article Access Map by Country/Region

1
Only visits after 24 November 2015 are recorded.
Back to TopTop