This blog is for all scientists (social or natural), not only for economists
Use of proper terms and definitions is very important in streamlining our understanding of the subject matter. Econometrics is particularly prone to misuse/misunderstanding of terms as the field is full of misnomers. In this blog, I will try to clear the meaning of as many terms as possible. Particularly, I will try to untangle the meaning of misnomers.
(One such misnomer in econometrics is the term “Random Variable”. I will talk about it in detail in future posts.)
Before we move ahead, let me talk about the name of the field itself! Professionals, students and professors alike have been using the term Econometrics to mean the use of quantitative (statistical and mathematical) techniques such as model specification, estimation, hypothesis testing and diagnostics in order to test a theory in economics or finance.
The term Econometrics is the sum of two terms Economics and metrics. So the term is made to give the flavor of economics and statistical measurement.
It sounds as if econometrics cannot be learned without knowing economics (can we?). Does this mean that the econometric tools are useful only in economics and finance? Is the knowledge of econometric modelling valuable for modelling phenomenon of ecology, physics, psychology or cosmology? The term econometrics explicitly favors economic theory.
Surprisingly, econometric modeling itself can be learned without any understanding of economics. The econometric tools we use do not know whether the phenomenon we are trying to learn about are from economics or physics. It is us, the users, who assign economic (or physical) meaning to the variables and the relationships involved. The tools themselves are secular in nature.
That means the econometric tools are basically statistical and mathematical in nature, which are universal in nature. The same tools can be used to understand any phenomenon of interest if the goal is to learn from the observable data about that phenomenon.
Now an obvious question can arise in our minds. If econometrics is just statistics and mathematics, why not call it statistics or applied statistics (statistics being a branch of mathematics)? Calling it statistics is just fine and accurate as well. But the term statistics basically gives the flavor of mathematical computation, theorems, lemmas, derivations and proofs, which are not among the main goals of scientists. The main goal is to learn from the data about the phenomenon of interest.
Therefore, we need a term that reflects our quest of knowledge about the phenomenon of our interest. And we want to achieve this goal by turning observable data into reliable evidence. This can be done using what are known as models. Models are simplified and yet adequate version of the reality (the meaning of simplicity an adequacy will be discussed in future posts). This adequacy part allows us to reliably learn about the infinitely complex reality.
So, I will use the term Empirical Modeling instead of Econometrics or Applied Statistics. This new term indicates that we want to learn about a phenomenon of our interest using empirical evidence via models. This term values both evidence based learning and model based approach. More importantly this term is not unnecessarily theory laden.
Hence, this blog (or empirical modelling) is not only for economists, but also for all other scientists (social or natural). In future posts, I will use examples not only from economics but also from ecology, cosmology, finance, health, management science and many other fields.
(In the next post, I will discuss whether a phenomenon is acquiescent for empirical modeling. Are there any phenomenon in nature or real life that are not acquiescent for empirical modelling?)
Niraj Poudyal, PhD
Reference
Spanos, A. (1999), Probability Theory and Statistical Inference: Econometric Modeling with Observational Data. Cambridge University Press. pp: 1-30.