There’s a funny point on which almost all economists that I’ve come across agree upon — from neoclassical to Marxian to Post-Keynesian. And that is that something which they call “nihilism” must be avoided at all costs. Let us first try to pinpoint what exactly this so-called nihilism is. A good starting point is Marc Lavoie’s seminal Foundations of Post-Keynesian Economics (a book which, incidentally, I hear is under substantial revision as we speak).
The second chapter of the book is entitled ‘Theory of Choice’ and deals with how Post-Keynesian economists should understand decision-making by economic agents. Lavoie is categorical in that marginal utility theory should be rejected and on that we fully agree. Tied to this Lavoie tells us that we should also reject the assumption of absolute rationality on the part of economic agents — again, I agree with this.
Next, however, Lavoie claims that we need to put forward what he calls a theory of procedural or bounded rationality. He quotes Cyert and Simon who define bounded rationality as such:
The rationality of the business firm is a rationality that takes accounts of the limits on its knowledge, on its information, on its capacity for computation, and on its understanding of theory. It is a rationality that makes extensive use of rules of thumb where a more exact application of theory is impossible whether because this theory is not understood, because the data needed for estimating its parameters is not available, or because the decision must be made under conditions of uncertainty. (Pp. 51-52)
I in no way disagree with this characterisation of how real world business firms make their decisions. I also think that this “rule of thumb” approach is applicable to some extent to individual agents — at least when it comes to activity such as meeting absolute living requirements (heating bills etc.); less necessary consumption and savings allocation/speculation have a completely different dynamic. I do, however, object to the use of the word “rationality”. We will come back to this in a moment but let us first turn to the question of nihilism.
Lavoie discusses what he calls “nihilism” on page 59 of the book. He does not provide a definition of the term, but I think we can derive one from what he has written. It seems that what worries Lavoie is that if we push the Post-Keynesian of uncertainty too far we have to then admit that there are no laws and regularities underlying economic activity — and thus that economics is largely a crock. Tied to this Lavoie is concerned that if we push the uncertainty concept too far we will find ourselves in a world characterised by total and utter chaos with no order whatsoever. What Lavoie calls “nihilism” then is more accurately called “epistemological nihilism” — that is a denial that we can have any knowledge of anything whatsoever.
He goes on to say that neither of these things follow because people follow rules of thumb in line with bounded rationality. Again, I agree with this. But the term “rationality” seems to me misleading and leads to the wrong conclusions.
The term “rationality”, as Joan Robinson once pointed out, is a circular concept. It assumes that agents act in line with some sort of theory — usually marginal utility theory — but the theory, in turn, is always grounded in the fact that agents are rational. Thus the theory really says nothing beyond “agents act as the theory says they act”. It is, in essence, a tautological argument.
The same holds for Lavoie’s use of the term. It doesn’t really tell us anything beyond the fact that agents act in line with how Lavoie thinks that they act. Thus when he uses the term “irrational” to describe the supposedly nihilistic universe of lawless chaos, all he is really saying is that in such a case the agents don’t act as he thinks that they act. In a very real sense then the term “rationality” is a moral argument in that peoples’ actions are judged purely on how well they conform to a given theory — neoclassical or otherwise.
This brings us to the next point: do economic laws really exist “out there”? I have argued before that they do not; that economics is simply one system among others which we use to organise our lives. In this conception it is true that some systems of action are better at achieving set goals than others. But the goals themselves need not be characterised as “rational” or “irrational” or anything else. They are simply seen as arbitrary and, ultimately, subject to our own moral judgement which precludes any application of economic theory.
I would argue that the “rules of thumb” that Lavoie discusses in his theory of bounded rationality are also arbitrary. Whether a firm or an individual has this or that goal in mind is basically arbitrary, although again there are better and worse ways of achieving this goal. This does not mean that these arbitrary rules of thumb lead to nihilistic chaos. They still produce regularities — although they certainly do not produce Laws, which in science must be timeless — and so we do not need to give up on economic theory altogether.
Still though, the specter of nihilism looms large and it is, so far as I can see, exaggerated. It is the yawning abyss that has people continuing to build bridges out of tautologies. Better, in my opinion, to just ignore it and get on with things. The day that economists stop using the term “rationality” is the day that they finally move away from becoming a modern day Church of Reason and become rather more humble in their role.
Fixing the Economists 
Nice post. Some points:
(1) They still produce regularities — although they certainly do not produce Laws, which in science must be timeless — and so we do not need to give up on economic theory altogether.
Yep, a great point which I agree with wholeheartedly. Economic empirical regularities are not “laws” in the physical sense, and to think so is absurd, the worst excess of neoclassical theology.
It’s like the law of demand, with its highly abstract ceteris paribus assumption, which cannot even be proven in its abstract form, as Steve Keen shows.
(2) “It seems that what worries Lavoie is that if we push the Post-Keynesian of uncertainty too far we have to then admit that there are no laws and regularities underlying economic activity — and thus that economics is largely a crock. Tied to this Lavoie is concerned that if we push the uncertainty concept too far we will find ourselves in a world characterised by total and utter chaos with no order whatsoever. What Lavoie calls “nihilism” then is more accurately called “epistemological nihilism” — that is a denial that we can have any knowledge of anything whatsoever.”
Actually, I do think Lavoie is right there.
But the solution is simply that Keynesian uncertainty really does come in degrees or grades, which are not quantifiable in an objective sense (say, in the way the probability of getting 3 in a fair game of dice is objectively quantifiable).
Inexact, unquantifiable uncertainty with unquantifiable probability comes with many inductive inferences, not just about the future, but about the past and present.
Don’t get me wrong. I agree with what I describe as Lavoie’s position in the above quote. I just think that he should stop using a theory based on the word (and concept) “rationality”. I think that because of this fear of nihilism which I think unfounded Post-Keynesians tend to fall back on the rationality concept. I don’t like this.
“I think that because of this fear of nihilism which I think unfounded Post-Keynesians tend to fall back on the rationality concept. I don’t like this.”
Buy in.
I am not an economist, so forgive my intrusion, but I find this very interesting.
I have never understood why economists use the term ‘rationality’ as though it is a fixed thing. Surely it is an evolutionary process – and it evolves from nothing.
For example, suppose I am given a task which I have never attempted before. The only possible strategy is to try things out randomly until I come up with something that works.
When I try the task for a second time the most effective strategy is to learn from the first attempt and to focus on strategies which worked the first time and avoid strategies which didn’t. I may still try out new strategies and/or variants of previous failed strategies but my second solution will be an evolution of the first.
On the nth iteration, I am likely to choose from a set of evolved optimal strategies, and recipes (or rules of thumb) for combining these strategies, with minimal additional experimentation.
Now suppose that, in the next iteration, you and I both attempt the task. My strategies and recipes are now very sophisticated while yours reflect my first random attempt. I will beat you even though we are both being ‘rational’. There is no fixed version of rational.
If you are smarter than me then, eventually, your strategies and recipes may evolve to be better than mine even though I started before you.
What have I missed? Why is this difficult? ‘Rational’ is simply an observation of behaviour at a point in time. ‘Irrational’ is just a failure to learn – like someone who can’t retain memories. The only ‘law’ is that strategies and recipes evolve from nothing through experimentation and learning.
I agree. But I think it goes a little further. “Rationality” in economics is quite clearly a value judgment. If you don’t act like economists think you should act then you are labeled “irrational”. Now consider the fact that most people do not act how economists think that they should act. Well, what does that make economists? I would argue that they are a priest caste.
“Rationality” is not the proper term for what the assumption actually involves. In this context it means that people, especially business people act pragmatically, that is, in their self-interest. People do not attempt to “maximize utility rationally,” i.e., in the same way. Rather, they attempt to optimize self-interest in light of limiting conditions and they do this not acting as isolated individuals but in the context of cultural conventions and institutional arrangements. Therefore, there are vast differences of behavior introduced by culture and institutions and a person position in society socially, politically and economically, e.g., wrt power.
The pragmatic assumption is significantly different from the rationality assumption in that the later is based on methodological individualism, which is a methodological assumption that has many problems empirically, in that humans are social animals that have developed societies as complex adaptive systems. Therefore, modeling human social behavior, of which economic behavior is only a part, as simple ergodic system, is naive. It seems to be chosen because it fits the market-view of society that favors privilege in a power structure, e.g, by rationalizing government non-interference in markets and promote the illusion that labor and capital are on the same level in bargaining power in the absence of collective bargaining by labor.
“Nihilism” is not the correct word, either. It should rather be called skepticism. A genuine scientists attitude is fundamentally skeptical in the sense of critical. This is what differentiates scientific method from apriori methods. While scientific models are hypothetical deductive systems, hence a priori, they generate hypotheses that are empirically testable, where the objective is to disprove the null case or show its probability of occurring in a stochastic approach.
So it is up to economists to show that they actually know what they claim to, and deductive methods alone don’t cut it in scientific enquiry.
“Law” is another misapplication in economics, which is social science that studies non-ergodic complex adaptive systems that are not law-like in the sense of hard sciences that study change in terms of motion. Change is social science is far more complex than motion.
Moreover, even in the hard sciences like physics, “law” as certain regularity is merely an analogy in that science is tentative knowledge that stands ready to adapt to new information. Science is the history of the growth and death of theories. Theories are replaced by more successful accounts when they get long in the tooth, with many ad hoc additions, as Thomas Kuhn has shown, for instance. Think epicycles. And who was the 19th century physicist that declared everything important had already been discovered, just prior to relativity and QM?
A further thought on “nihilism.” I’ve noticed that orthodox economists routinely reject heterodoxy as not have a model. When asked, they say that without a (deductive) model, economics would be no different from sociology or other soft “sciences.” So they stick with models that are non-representational since they have physics envy. One can hear the rationalizing that after all, QM is stochastic and sub-atomic phenomena are not actually observable other than through assumed effects; it’s purely theoretical.
This is like the drunk looking for his lost keys under the lamppost even though he dropped them on the lawn “because there is more light there,” as others have pointed out.