چکیده:
During the 1960's, many as was firmly supported by the
historical founders of econometrics, had hoped that
econometrics would provide a sound scientific foundation for
econometrics in which each element of specification would be
determined primarily on the basis of economic theory. However,
due to misusing of econometrics and also wide usage of the so
called cookbook econometrics, many researchers - especially
economists - have recently lost some confidence in econometrics.
The present paper discuss the ingredients of the scientific
econometrics and suggest ways of helping to renew optimism
about scientific econometrics or as I have called it
'scienceometrics” with an emphasis on the ethical issues.
خلاصه ماشینی:
They have manipulated the data (using a specific sample of sub-sample), the variables (using different versions), lag-structures (using different time-lags), mathematical forms of the model (linear or non-linear), estimation procedures (weighted instead of ordinary least squares), and so on without any scientific justification and then to choose the equation that conforms the most to what they want the results to look like.
This way of dealing with econometrics and model selection which may truly be called "cookbook" econometrics is completely opposed to the scientific econometrics in which each element of specification should be determined primarily on the basis of economic theory rather than on the results of a desired estimated regression equation.
It should be noted that from the scientific econometrics point of view the best models arc those on which much cares has been spent to develop the theoretical underpinnings and as mentioned by Studenmund'" only a short time is spent pursuing alternative estimations of that equation.
Even causality tests such as those introduced by Granger and Sims and others which are based merely on statistical and mathematical reasoning rather than economic justification, can only be used to show the direction of a causal relation and in scientific econometrics I suggest these tests should be used as a last resort regarding the direction of the relationship when there are ambiguities".
Based on discussion given in this paper the following are some important suggestions to improve the renewed optimism about applying scientific econometrics: 1) Economists should use structural models that are characterized by the use of economic theories in specifying relations rather than the vector autoregressive models, which are really measuring without theory.