The Test of Science in Economics

Michael Hudson writes in J is for Junk Economics (2017: 333) that:

When criticized, the authors of... [the mainstream] textbooks [of modern neoclassical economics], from Paul Samuelson to Bill Vickrey, say that it doesn't matter whether economic theory is realistic or not [the Milton Friedman line]. The judgement of whether economics is scientific is simply whether it is internally consistent.

Thus, for modern orthodox neoclassical economists, the criterion of science in economics is not the underlying realism of its economic theories but their internal logical consistency.

Now, with this sort of logic, anything could be deemed scientific, such as an economic theory about the price phenomena of a market economy like capitalism based on the hypothesis there are fairies in the garden 'crying' out like a grand auctioneer the market prices of all commodities. A simple logically deductive example can easily establish this point.

Fairies in the garden 'cry' out like a grand auctioneer the market prices of all commodities (major premise).

The market prices of all commodities are signals for all markets to clear at a general equilibrium price (minor premise).

Therefore: 

The 'cries' of fairies in the garden leads to all markets clearing at a general equilibrium price (conclusion).

This is, logically speaking, an internally consistent economic theory about how all markets clear at a general equilibrium price because of fairies 'crying' out in the garden like a grand auctioneer, as the conclusion deductively follows from its major and minor premises (in symbolic terms: if A is B, and B is C, then A is C). But, a moment's reflection is enough to tell us that we should reject this economic theory as it's based on an unrealistic hypothesis: there are no fairies in the garden actually 'crying' out like a grand auctioneer the market prices of all commodities: this is a piece of fiction! Yet, despite this glaring lack of realism, we are still expected to accept this theory as being scientific, all on the grounds of it being internally consistent - the criterion of science for modern orthodox neoclassical economists, according to the likes of Samuelson and Vickrey.

Now, this is where the influence of Milton Friedman's views about methodology (his F53 paper) come in. His methodological views are meant to justify the construction of patently unrealistic theories in modern orthodox neoclassical economics and fend off any criticisms about their unrealistic nature.

It basically does this by invoking the 'as if' mantra. That is, while it may be accepted that the economic theory in question is unrealistic because it's based on an unrealistic hypothesis or a set of unrealistic assumptions, that doesn't matter since it's enough to just treat the underlying hypothesis or assumptions of the theory in question as if they're realistic (or true) - not that they really are. Thus, in the particular above theoretical case of how all markets clear at a general equilibrium price, we should look at its underlying hypothesis or assumptions as if they're true in the sense that there really are fairies in the garden who 'cry' out like a grand auctioneer the market prices of all commodities, even though we know there are no such things in reality.

But why should we do this? Well, for Friedman, the short answer is that it enables one to 'scientifically' draw the right sorts of conclusions that one is after in their economic theories about the economic phenomena of a market economy like capitalism. (NB. I'm only giving the guts of his methodological views here.)

And furthermore, for Friedman, the only real scientific test of any economic theory is whether its predictions are confirmed by the world of experience. If a theory's predictions are discomfirmed by the world of experience then the theory is to be either rejected or revised until its predictions are finally confirmed by the world of experience. Equally, if the predictions of a theory are immediately confirmed, then the theory is immediately accepted. Either way, for Friedman, the realism of the underlying hypothesis or the assumptions of the theory in question is irrelevant as to whether it's accepted as a valid 'scientific' theory in economics.

However, as pointed out in an earlier blog post - "Scientific Truth Matters" - it's not correct to say that the realism issue is irrelevant as to whether we accept a theory as being scientifically valid. As I argued there, the more realistic a theory is in terms of its underlying assumptions, the more accurate its predictions are likely to be, as exemplified by Einstein's general theory of relativity - which had come to ultimately replace Newton's theory of universal gravity as being a more accurate scientific representation of our solar system and universe because of its more accurate or truthful assumptions about the causes and nature of gravity.

It's notable, by the way, that the central criterion for determining whether a theory in economics is 'scientific' is different for both Friedman and Samuelson-Vickrey. For Friedman, as we've just seen, the criterion is whether a theory's predictions are confirmed by the world of experience; while, for Samuelson-Vickrey, as we saw earlier, the criterion is whether a theory is internally consistent. Despite these differences, what they all accept however is Friedman's original assertion that the issue of realism is irrelevant to the acceptance or non-acceptance of the 'scientific' status of a theory in economics; hence, the influence of Friedman's methodological views on the economics profession, especially its modern orthodox neoclassical economics school (the dominant school of modern mainstream economics). However, as I've just pointed out, such a methodological view is quite questionable. Hence, it's not a sound or firm basis for resting their respective criterion of 'science' on, in economics.

At any rate, here's one pivotal example (as provided by Hudson) of how modern orthodox neoclassical economists try to demonstrate the 'scientific' status of their type of economics on the underlying basis of Milton Friedman's methodological views in modern mainstream economics.

It needs noting that on the subject of Milton Friedman's methodological influence in modern mainstream economics, one of the key issues to tackle is his 'as if' approach in economics, as this is really at the centre of not just his own methodological views but also at the practice of economics itself within the theoretical framework of modern orthodox neoclassical economics. In doing so, as I shall show in other blog posts on this topic, we will need to confront Friedman's 'as if' approach with the Galilean method of abstraction and idealisation in science. This is the scientific practice or 'mental act' in science which seeks to, on the one hand, leave things out while giving a literally true description of something (the method of abstraction), while, on the other, treating things as having features they clearly do not have (the method of idealisation). Such a method of science is rooted in reality, not independently of it, as in the case of Friedman's 'as if' approach (see my blogpost, "A Follow-Up: a Conceptual Distinction"). Therefore, such a confrontation will reveal that Friedman fails to understand some essential features of how actual science works in a methodological sense, as it goes about its central task of revealing the hidden essence of the phenomena of the world and consequently producing as practical as possible an accurate representation of the phenomena of the world.

Also, as we shall eventually come to see, Milton Friedman's methodology is best construed as being a form of instrumentalism - one of the major outshoots of modern positivism in the philosophy of science, and which also stands in opposition to the views and aims of scientific realism (meaning, it rejects any theory which claims to have discovered the hidden essence of phenomena on the philosophical grounds that they are unknowable).

Addendum

Milton Friedman's recommendation that economists should ignore the realism issue is a way of sidestepping it rather than addressing it.