Neoclassical economics as a utility function that represents a consumer's preference ordering over a choice set.

Joan Robinson criticized utility for being a circular concept:

"Utility is the quality in commodities that makes individuals want to buy them, and the fact that individuals want to buy commodities shows that they have utility"

This criticism is similar to that of the philosopher Hans Albert who argued that the ceteris paribus conditions on which the marginalist theory of demand rested rendered the theory itself an empty tautology and completely closed to experimental testing. In essence, demand and supply curve (theoretical line of quantity of a product which would have been offered or requested for given price) is purely ontological and could never been demonstrated empirically.

In my understanding, the neoclassical economics is based on utility. So if this concept is no testable, and circular, how neoclassical economics can explain anythings?

I'm not saying the system can not predict anythings, because if we record a lot of behaviors, we can certainly predict some behavior by comparing similar patterns (it's mainly the logic behind machine learning). But we should not call that a theory neither an explanation.

I end up with those interrogations:

Related content

As I'm exploring this question, and will add some content I stumble upon that I think add some context.

  • 2
    $\begingroup$ Simple syllogism: 1) If the concept of utility is not testable and circular, neoclassical economics cannot explain anything. 2) Neoclassical economics can indeed explain something. 3) Therefore, the concept of utility is either testable, or non-circular, or both. $\endgroup$
    – Herr K.
    Commented Nov 13, 2018 at 16:40
  • 1
    $\begingroup$ If you give people a set of goods and ask them to rank them in how much their desire them, they will be able to do it easily. They may not be able to explain exactly why they made the ranking they did, but they can do it. So utility exists at the personal level even if we may have trouble articulating it. $\endgroup$
    – zeta-band
    Commented Nov 13, 2018 at 18:39
  • $\begingroup$ @HerrK. to be a syllogism, you need 3 categorical propositions. I'm not an expert in logic but the first one doesn't seem a categorical proposition but a conditional statement. $\endgroup$
    – gagarine
    Commented Nov 13, 2018 at 19:47
  • $\begingroup$ @gagarine: Only in Aristotelian logic are syllogisms restricted to contain only categorical propositions. For instance, hypothetical syllogism and disjunctive syllogism both contain non-categorical statements. The name of the argument notwithstanding, I think what I said was both valid and sound. $\endgroup$
    – Herr K.
    Commented Nov 14, 2018 at 1:32
  • $\begingroup$ @HerrK. I see, thanks for the precision. So what kind of syllogism it can be? If you can formalize it, that can be an pretty good answer. I tried but I was not able to do it :/. $\endgroup$
    – gagarine
    Commented Nov 14, 2018 at 11:45

2 Answers 2


Let's be clear--the concept of utility is both unfalsifiable (as a singular proposition) and useful.

Joan Robinson is right, of course: Utility as a definition is circular. If you read a textbook that discusses the foundations of economic theory (such as the popular graduate text by Mas-Colell, Whinston, and Green), then you start not with utility, but with preferences. You would characterize an individual as being defined by a "preference relation" which describes, for every pair of potential allocations (i.e., sets of goods), which they prefer.

Some (not all!) preference relations are rational, which means that they are both complete (you have a well-defined preference over any pair of allocations), and transitive (if you like A more than B and B more than C, you like A more than C). A subset of those preference relations can be represented by a utility function (again, not all--the classic example is lexicographic preferences).

So a utility function is really just a thing that represents some underlying preferences. The statement "Steve's preferences are accurately modeled by a utility function" are not restrictive enough to be tested in any practical way. Rationality in general is hard to test as well, as you can always explain away seeming rationality violations with unusual preferences.

Instead, you should think about utility as a framework, not a theory. It's a language for developing theories. For example, you can describe Friedman's Permanent Income Hypothesis without invoking utility at all, but describing it in terms of utility is a lot clearer and simpler!

As an aside, the blog post you linked to is utter tripe, from a guy who seems to love bashing economics while not actually knowing all that much about it. I would avoid him like the plague.


Ranked order preference within effective demand does have problems; but not to my mind the problem of tautology. It is a model, one which reduces human conduct to the terrain of ranked order preference within effective demand. “Homo economicus” is a useful stand in, especially given the problems of utilitarianism in terms of the social reconciliation of desires and desires being unrestricted by effective demand. The concept of utility maximisation in homo economicus is a similar limitation.

Establishing a major field of social science on an assumption, or a simplified model, isn’t to my mind a bad thing. Especially as it is an up front assumption for the purposes of building models.

The theory has falsifiability?

Falsifiability is one conception in the philosophy of science to help with questions of what science is. Other philosophers of science have suggested it is of limited usefulness (Kuhn, Lakatos, Feyerabend.)

Given that marginalisms have staked a great big place in the academy, the state and the market I don’t think it is particularly useful to beat them with the stick of falsifiability, or even ad hocery; Tempting as it might be to use Popper on marginalism given Popper’s work on Marxism. Science generally doesn’t do “True” thanks to Hume. Falsifiability is one way forward, but open to the claim that falsifiabilty itself fails to adequately reflect empirical life.

Philosophy of Science and History of Science and Science and Technology Studies have all come around to the question of “isn’t it just all one damn thing after another?” And to a certain extent that is it.

But we should not call that a theory neither an explanation.

Theories and explanations as great edifices aren’t necessary. Doctoral students, Senior Lectureships, ARC Grants, and honours students appointed to Treasury or Finance are. Expecting Economics as science to be more coherent than Physics, or Mathematics, is potentially admirable but actually heart breaking. My only suggestion would be avoiding shouting “the peasants wear no clothes,” at the heterodox who attempt to maintain internal consistency, empirical reflection and 3rd party impact factors, when the emperor of thinking about human production and exchange is also nude.

Science is hard. Field specifying assumptions are often inadequate. Most scientists ignore the problem and get on with churning out papers—some even believe their research has utility.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.