post-autistic economics
media archives







Global Economic News


Paecon Home Page

PAE Media Archives Index

PAE Movement News

PAE Newsletters



 
A rejoinder to James K. Galbraith
and a contribution to the ongoing debate


Jacques Sapir  
(L'École des Hautes Études en Sciences Sociales, Paris)


In the post-autistic economics newsletter, issue no. 4,  James K. Galbraith wrote a very interesting and convincing reply to Robert Solow.  In the same issue, Joseph Halevi aptly described what he calls a "Franco-American neoclassical alliance", pointing out that Blanchard's reply to the students’ manifesto was dogmatic and hollow.  I would like to jump into the fray and to offer some points in support of a post-autistic approach to teaching, learning and doing economics.

 

Before going to the main points, I want to say that I completely agree that teaching (and learning) is the basic issue and that to my knowledge, even if some professors  have appealed against the student’s views, nothing similar has appeared from the students’ side.

 

My experience in teaching economics began in the late 70's. After spending 4 years as a high-school (Lycee) teacher, I have taught macroeconomics at the University of Paris-10/Nanterre (where the department of economics is probably one of the least autistic in France), before moving to the École des Hautes Études en Sciences Sociales, a specifically  post-graduate institution. I have also taught regularly in Moscow (at the Higher School of Economics and at Moscow University), and taught many seminars in the USA, Italy and the UK.

 

What French students described in their manifesto is the plain truth and the result, not of France deviating from "good economics" as taught in the USA, but, to the contrary of what Robert Solow suggested, the French system converging with the US one.

 

Now, what is the discussion mainly about? The focus is on the link between teaching and science and the point is realism vs. axiomatism. One of the most important issues raised by this movement is the very fact that you can not separate tools and contents, teaching and what you are supposed to teach. By engaging the battle on this field, the movement already won a major epistemological victory. But, if even Solow acknowledges that the discussion is not about mathematics in economics, clearly he does not grasp the meaning of the "realist" requirement. I, however, suspect that some of my post-autistic colleagues do not either.

 

The real issue is not to know if the realistic content of economics is to be increased just slightly or in a more meaningful way; rather it is to know if realism is a central requirement or not. This leads to an old but forgotten debate, the one about "natural" or "exact" laws governing economic activities. Without going back to Carl Menger, we have to remind ourselves that if there is something like such laws, then realism is not a requirement for theorizing and a quest for realism not only is not needed but could even prevent us from reaching a complete understanding of a system obedient to laws.

 

Now, what has usually been forgotten by people supporting the axiomatist approach, and this is fairly obvious with Solow and Blanchard, is the type of conditions required for "natural" or "exact" laws to exist in economics. Either there is to be no creative interaction between economic agents (Robinson Crusoe on his island before the landing of Friday), or an agent's behavioural patterns are to be context independent. The first condition implies there is no society and no economy. The second one is a key assumption of neo-classical economics (preferences are supposed to be context independent, transitive and continuous).

 

Unfortunately for our neo-classical colleagues, repeated tests have amply demonstrated that preferences are context-dependant (Amos Tversky's framing effect) and are neither transitive nor continuous (See Kahneman, Slovic and Lichtenstein works). Unless these tests have been proved to have been faked or incomplete or wrong in any given sense, the very Popperian methodology our colleagues are so fond of should have led them to delete this key assumption about behavioural patterns, with the obvious consequence that there is no "natural" or "exact" laws in economics.

 

It is a well-known fact that very few of them, Kenneth Arrow being the best known exception, are willing to accept this conclusion. If they were, they would understand why realism is a central requirement in economics.  Therefore, it is important for non-autistic economists to understand that it is not just "more realism" that they must fight for, but also, and more importantly, for a methodological posture revolving around the requirement for realism.

 

This last point raises another problem, the all too frequent confusion about the level of abstraction in economics. Having worked for many years on the Soviet, then the Russian, economy, maybe I have developed a certain sense on that matter. Going back to James K. Galbraith’s contribution, to discuss the market for low wage labor as a specific case of the labor market is not the same thing as to discuss the market for low wage labor in a given country, at a given time and with the given set of institutions governing relationships between the internal market and the world economy.

 

By the same token, when we discuss the labor market, the word ”market” does not have the same meaning as when we use it in a discussion about relative pros and cons of market and central planning.  Another extremely dangerous way to do economics is illustrated when Robert Solow, in his Le Monde article, switches from an highly abstract level to directly operational issues as if concepts were part of real life. Spinoza wrote that the concept “dog” does not bark, but nevertheless, if we understand how to use concepts, “dog” is still useful for understanding life where dogs bark and even bite. Realism is not to be confused with reality. The fight against the misuse of axiomatism in economics (and generally speaking in the social sciences) is not an anti-theoretical turn but a different approach to theoretical thinking.

 

Having gone so far in support to the post-autistic approach I must confess some unease about the widespread use of the term “pluralism” in the PAE-Newsletter. First, it is to be understood that I support teaching various theoretical approaches in economics if done in a critical way. Second, I also support the idea that a scientific debate must be done in an open-minded way and particularly when it concerns academic and professional journals. But, and this is no small caveat, no scientific debate is possible without agreed rules about what is correct and what is not. No geographer would teach her/his student the flat Earth theory except as an historical example of false geographical understanding of the world. No historian of the Soviet Union would now teach her/his students the 30's without talking about the 1932/33 starvation, whatever explanation he/she could give about it. Anyone pretending in the scientific community that Earth is flat and Soviet peasants were rich and happy in 1933 would be called a faker or a lunatic. So there are bounds to pluralism.

 

Reclaiming pluralism makes sense if and only if we develop a scientific methodology consistent with the requirement of realism. I understand that among post-autistic economists some still believe that the Popperian approach is worth some consideration. I have my doubts about it if only because the logical contradiction embedded in Popper's falsification theory, when applied to social sciences, as demonstrated by the late Professor Quine. It's not the place here to fully address this issue but my position is that without understanding all the realist methodological implications, pluralism could become perverse, that is the simple addition of conflicting points of view without any means to critically assess them, or even the understanding that some are not compatible with others. Otherwise we will run into big trouble, not least the one of theoretical inconsistency, and offer a weak flank for autistic colleagues to attack. The so-called patchwork effect I am afraid of is already pervasive among mainstream economists.

 

I would not spare PAE-Newsletter readers this short account of something that happened to me years ago at the IMF. After quite a long meeting with people of the Europe-II department, and a very heated debate about what was really happening in Russia, the chairman of the department wanted to make a point. He then very solemnly declared: "Remember, inflation is,  every time and everywhere, a macroeconomic phenomenon, as demonstrated by von Hayek". Oooops! You could believe in standard macroeconomics, or believe in Hayek, but the fact is that the whole Hayekian understanding of inflation is grounded on a microeconomic process. You can't mix both.

 

To go back to the issue of individual preference and how economists are to write their model specifications, how could we accept context-independent, transitive and continuous preference functions when we know it's wrong? Daniel Hausman in his 1992 book, The Inexact and Separate Science of Economics, wrote a full chapter about how mainstream economists consistently refuse to acknowledge the preference reversals phenomenon, and called that dogmatism. If autism is to be described as an acceptance of fantasy rather than reality and a withdrawal from reality, then nothing can better describe some of our colleagues’ behaviour about the theory of individual preference. But to separate ourselves from autistic economists we then have to elaborate specific assessment rules as to organize scientific pluralism.

 

This leads me to a short comment about the French PAE curriculum, a comment I made at the October 2000 meeting in Paris. It's very important to re-introduce History of Economic Thought into the core curriculum. It would be even better if backed by the introduction of lectures in methodology and epistemology. If economics belongs to the social sciences, then students, who are to be taught economics, deserve to learn what is the methodology and epistemology of social sciences. I know that students usually don't like methodology and epistemology, mostly because philosophy as taught in secondary school is either boring or has become a complete ideological playground. What's more, the jump to the level of abstraction needed to understand what methodology and epistemology are about is distinctively at odds with the now fashionable posture where teachers and professors are supposed to explain anything from the pupil/student's direct experience point of view. This fallacy now dominates pedagogical practice in French education. Still, I believe that if we could cut by a half the time devoted to math in the curriculum and to substitute a similar amount of time to methodology and epistemology, everybody would be better off.