Neoclassical Economics and Domain Assumptions

How the learned practice sophistry

by Joe Blackwell

A recent paper authored by two mainstream neoclassical economists [1], Salomon Faure and Hans Gersbach, has caused quite a stir amongst prominent heterodox economists such as Steve Keen and Richard Murphy. The paper seeks to compare and contrast two opposing models of commercial bank lending behaviour (endogenous money model and loanable funds model).

According to the authors,

“Almost all models of banking—be they micro- or macro-oriented—are based on the so-called “loanable-funds approach to banking”: Banks are financed through deposits, equity, and other financial contracts, and then they lend to firms or buy assets. In our current monetary architecture, however, the opposite process is at work. Banks start lending to firms and simultaneously create deposits.”[1]

In other words, the loanable funds model posits that banks are merely intermediaries that lend out the deposits of savers, this leads to the conclusion that banks do not create new spending power when they lend but merely redistribute the spending power of savers to borrowers. As the authors explicitly admit, this model underpins nearly all neoclassical micro and macro models that feature bank lending and is a completely inaccurate description of reality. Endogenous money is far closer to the reality of banks’ lending behaviour. This model holds that rather than lending out the deposits of savers, banks create deposits on their own ledgers when approving loans. This therefore leads to the conclusion that bank lending increases the spending power of the economy and actually increases the broad money supply. This therefore means that all the neoclassical models that ignore lending in the aggregate, (for they claim funds are intermediated from savers to borrowers, so this doesn’t increase Aggregate Demand) are excluding one of the key dynamic mechanisms in the economy.

You would assume then that neoclassical economists would happily throw loanable funds theory to the dustbin of history only to feature as a curiosity in the history of economic thought lectures. You’d be wrong. Instead of behaving in a scientific way, neoclassical economists have instead decided to protect the base theory by imposing new assumptions. In this case,

“We establish a benchmark result for the relationship between the loanable funds and the money-creation approach to banking. In particular, we show that both processes yield the same allocations when there is no uncertainty and thus no bank default. In such cases, using the much simpler loanable funds approach as a shortcut does not imply any loss of generality.”[1]

So loanable funds and endogenous money lead to exactly the same lending behaviour as long as we assume that uncertainty and bank default doesn’t exist, and therefore the two are interchangeable so we might as well use the much simpler loanable funds model.

This fallacious use of assumptions is a common hand waving technique employed by neoclassical economists as they conflate two different types of assumptions. The two types of assumptions relevant here are simplifying assumptions and domain assumptions. Neoclassical economists claim that all their assumptions are simplifying assumptions. The common analogy given in the popular micro textbooks is,

“By a model we mean a simplified representation of reality. The emphasis here is on the word “simple.” Think about how useless a map on a one-to-one 2 scale would be. The same is true of an economic model that attempts to describe every aspect of reality.”[2]

This seems intuitively reasonable. If economic models tried to capture every aspect of reality, the models would become so complicated that identifying the causation between different economic variables would become impossible. The problem is that mainstream economists don’t just use simplifying assumptions, they also use domain assumptions that are then labelled as simplifying assumptions.

A simplifying assumption is an assumption that excludes some aspect of reality to try and simplify the model so it can focus on the specific phenomena that is being studied. An example of a simplifying assumption would be a model of the orbit of the earth around the sun that ignores the orbit of other planets. The model can then be compared to the data to see if the orbits of other planets play a large enough role in the data generating process in reality – if not, then models that study the orbit of the earth around the sun can exclude the orbit of the other planets in the solar system.

A domain assumption on the other hand is when the modeller makes a fundamental assertion about the domain of the model that is untrue in reality. For instance, if we modelled the solar system as if the sun orbited the earth, then the model might give some insight into a solar system where that was true; but it says nothing about the solar system we actually live in. The same logic applies to many of the assumptions of neoclassical economics, including the assumption made in the aforementioned paper. To assume that uncertainty doesn’t exist isn’t merely a simplifying assumption, it’s a domain assumption. Therefore the findings of the paper are only true in a world where uncertainty and bankruptcy don’t exist. In reality, uncertainty is an easily observable phenomenon, so the findings of the paper don’t apply to reality.

When economists try to hide a dodgy assumption by claiming it’s merely a simplification, it is likely that their assumption is a domain assumption, which discredits their whole chain of logic.

[1] Faure, S. and Gersbach, H. (2017). Loanable Funds vs Money Creation in Banking: A Benchmark Result.

[2] Varian, H. (2003). Intermediate microeconomics. 8th ed. New York: W.W. Norton & Company, pp.1-2.

Joe Blackwell is a London based economics student with an interest in MMT, Post Keynesianism and critiques of orthodox economics. His undergraduate background was in political science – the effect of economic conditions on the rise of populism in Europe and the USA. Joe can be reached over Twitter: Email:

One thought on “Neoclassical Economics and Domain Assumptions”

  1. The issue with the model is in my opinion far more complicated. They assume a pre-allocated endowment of a single good and claim that the distribution is identical with and without endogenous money (if there is no uncertainty and bank defaults). Well this is hardly surprising if we understand the original Wicksellian model based on natural rate of interest when there is no inflation, the system with endogenous money is in global equilibrium and behaves as-if the capital was lent in kind. So the problem with the current model is that it is not demand-constrained as there is an endowment emulating the global equilibrium and Say’s law, not that risk of default is ignored. We know that any Post-Keynesian model if pushed to the productive capacity limits will only generate inflation if more money is spent. Then the loanable funds theory applies to the world of endogenous money. Want to invest more – have to save more. The difference between the LWT and endogenous money models only becomes visible if the system is demand-constrained and operates at a certain bias point set by the average propensity to consume out of (expected) disposable income equal to 1 (the Keynesian cross), with spare productive capacities. Then and only then lending creates ex-nihilo extra spending power and affects the real economy because it shifts expected disposable income (assuming no changes to the consumption schedule), money neutrality no longer exists. Investment and debt-financed consumption drives profits and saving as explained by Kalecki. So my summary of the paper is “nice try but no cigar”. We still live in a Kaleckian world of endogenous money…

Leave a Reply

Your email address will not be published. Required fields are marked *