Methodology in Economic Science

Chapter 1, Section 8: Methodology in Economic Science

If you don't understand something in this chapter, don't take that into heart. The methodology of science involves logic and the theory of knowledge that come from philosophy. These are almost the most profound knowledge that human have ever gained throughout history. Thought I took lessons from masters, I don't think I know that much; if I use only simple terms, I am afraid the ideas may not get correctly conveyed. The methodology of science is so profound that sometimes the masters of logic may even not agree with each other; the good news is one's accomplishment in science doesn't depend much on this learning. Giants of science that don't know much about the methodology are uncountable; conversely, experts in the methodology are rarely eminent scientists. Logic usually goes to the extremes of ivory tower. Its subtleties sometimes are marveling, but to reach perfection one has to pay a very great price.

From the point of rigorous philosophical logic, what I know is only very general. I studied this knowledge fifty years ago. However, there is another angle the methodology can be viewed, the empirical methods that function in the transition between abstract theories and the real world. About this aspect I know much more. The content of this chapter combines both philosophical logic and empirical transition, so it's quite different from those in other books. Anyway, practical sciences must walk out of the ivory tower in the end.

I placed the Methodology of Science at the head of this book and used an entire chapter for it, is not because the knowledge is indispensable for this book. More important is, the Chinese cultural tradition only talks about benevolence, righteousness and virtue; it lacks any empirical character, and has a mistaken attitude towards science deep in its heart. The Three Principles of the People and Marxism, or any other isms, that have had great influence on the Chinese in the twentieth century, only add another opaque film to Chinese's understanding of science. It's been stated that this book is mainly for Chinese students. In my view, the methodology of science is more important for the Chinese than for other western nations. Yet never enter the blind alleys in methodological logic! A general idea is already sufficient.

Economic explanation is an empirical science. In essence it's the same as other natural sciences, and both use the same methodology. However, economics differs greatly from natural sciences in the characteristic of its subjects, and focal methods recruited are different as well. On the subject part, the difference is for two reasons. On one hand, not like in other sciences, the lab of economics is the real world instead of something built and operable by economists, so the difficulty in observation is exceptional. On the other hand, economics is the explanation of human behaviors, but economists are also human, so the explanation is inevitably reading themselves as well. Therefore making objective judgments in economics is usually much harder.

As for recruited methods, the emphasis in economics differs from that in natural sciences in the following aspects. First, I don't think economics should bring in that much influence by physics. As is mentioned previously, the so-called equilibrium and disequilibrium are real phenomena in physics, but in the world of economics there is no such real stuff and they are only concepts, which I have explained. Of course, there are those economists who think equilibrium and disequilibrium are observable market phenomena. That is embarrassingly big mistake. Students please pay attention, when I mention equilibrium in Economic Explanations, I mean a refutable theory is established because well enough constraints have been installed, and I'd never intend any observable steady point.

I don't think mathematics should be that heavy in economics as well, although today's journal papers in economics use more maths than in physics. Except physics, other natural sciences don't use maths that much. I don't mean maths is of no use in economics, it's just maths is never economics. Maths is an amazing language: anything obtained from formula transformation immediately follows logic. Nevertheless, the correctness in logic doesn't guarantee correctness in content. There are indeed those who are efficient at thinking with formulas, but for me it's just adding another layer of restrictions and keeps me from free thinking that I am more capable of. Without maths, my reasoning logic has never gone wrong. I suggest you students learn more maths, but when one thinks, he must take into consideration which his mind is better at, maths or direct reasoning. With enough practices, reasoning without maths can be much freer and allows for more imagination.

There is a common mistake. Lots of people think reasoning with maths or statistics formulas can be more accurate. This is not correct. Measuring is arranging numbers, while accuracy is the public recognition of this arrangement. (Translator: I don't understand either.) This should be another topic in philosophy, and I will demonstrate how to handle it when discussing transaction cost in Volume 3.

Second, I have never dealt with hypothesis test in natural sciences, yet hypothesizing and validation in economics always start from the variation of constraints, i.e. test conditions. When the happening of A implies that of B, it's actually saying the variation of A leads to the variation of B.

As is said, prediction is the same thing as explanation, it's just one is beforehand and the other afterwards. Prediction is to have the variation(s) of constraints in mind first and then infer what phenomenon is going to happen; explanation is to watch a phenomenon happen first and then go back to check what variation(s) of the constraints causes it. Same logic structure, isn't it? Yet dealing procedures of the two are different, which practice is harder is not easy to decide. Just imagine, when facing a phenomenon, to explain it, we need find the variations of constraints that cause it, but within infinite number of constraints, which one or combination shall we choose? Without doubt, backtracking the variations of constraints requires the guidance from some theory. Not easy. Then how about prediction? We may see a constraint has began to change, and then predict by theory what phenomenon is going to happen. The problem is, the variation currently revealed may not be stable and keep changing from time to time, that can easily void the previous prediction. In 1981, based on some changes of constraints I believed stable enough, I predicted that China was going to take the way of market economy. Yet it came true is all by God's blessing, as nobody guaranteed the stability of those changes would continue.

So here comes another related important topic. To make any beforehand prediction or after explanation, one has to master well enough the assisting theory and the application of concepts. However, all the books I have seen that claim to be application of economics usually just throw out a theory, grab some real world examples and then press them in. Basically this is just seeking the right, and violates the dogma of scientific validation: to strive for "wrong" but never get refuted, i.e. reputability. Being refutable and being testable is pretty much the same. As those application books have a different consideration and purpose, it's quite improbable that we can learn much from them.

The last thing I need to tell is, anything invisible is unable to test. This is philosophically very simple. Yet today's economics violates this principle frequently and that results in a catastrophic development. When asserting if A happens then B happens, we must make sure both A and B are real facts, or at least in-principle observable or touchable, or else its validation would become impossible. The starting point of a theory is usually abstract, that is to say, there exist some variables not even in-principle observable. There is one more ghost to service when one more censer is added, so we need avoid unobservable variables as much as possible. After years of exploration, there lefts only one unreal variable that I can't avoid and have to accept: quantity demanded. Only for this one had I gone over many difficulties.

In today's development of economics, there are so many unobservable variables or behaviors: game, motivation, shirking, blackmailing, threatening, concealment, laziness, opportunism, etc. I am not saying there are no such things, it's just they are not observable or measurable and thus not testable. When you can't test something due to its invisibility, you may still tell stories, and the stories can follow logic and sound trustable. Kind of like religions. A science impossible to be tested cannot explain anything. (Translator: there is a fallacy here that, the writer has confused the concepts abstracted with the variables to be tested. For example, why would we ever need to test the existence of a game? What we need to test is only the result, or the prediction, of a game. In this sense, concepts like game are pretty much the same as the "quantity demanded" that the writer has retained.)