I’ve been meaning to write about this topic for some time, but a Tweet the other day reminded me that there was still major misunderstandings of what Modern Monetary Theory (MMT) represents and that it was time to clarify some of those errors in comprehension. Specifically, there is a current out there that considers MMT to be incorrectly labelled because according to the argument there is no theory involved. It’s hard to imagine why anyone would think that but the fact that they do tells me that I should write this blog post. As I noted yesterday, our Macroeconomics textbook to be published by Macmillan Palgrave in February 2019 is full of theory. It has a lot of description, taxonomy, accounting, history, and philosophy, but also a lot of theory that ties some of those other components together in a meaningful way. The T in MMT is not a misnomer. The Tweet I saw the other day also said there was nothing new in MMT so what’s with the modern bit! I have already dealt with that issue in the past.
On the question of the ‘modern’ bit, the following blog posts are relevant among others:
1. Modern Monetary Theory – what is new about it? (August 22, 2016).
2. Modern Monetary Theory – what is new about it? – Part 2 (long) (August 23, 2016).
3. Modern Monetary Theory – what is new about it? – Part 3 (long) (August 25, 2016).
When I don’t know the meaning of something I initially look up the dictionary, which in part, also helps me understand the etymology involved.
The Oxford Dictionary entry for Theory is:
1 A supposition or a system of ideas intended to explain something, especially one based on general principles independent of the thing to be explained.
1.1 A set of principles on which the practice of an activity is based.
1.2 An idea used to account for a situation or justify a course of action.
1.3 Mathematics A collection of propositions to illustrate the principles of a subject
It comes from the late C16th “denoting a mental scheme of something to be done) – “contemplation, speculation”.
The Oxford Thesaurus notes:
hypothesis, thesis, conjecture, supposition, speculation, postulation, postulate, proposition, premise, surmise, assumption, presumption, presupposition, notion, guess, hunch, feeling, suspicion.
This list isn’t very helpful because it conflates many different conceptual structures with the work theory.
For example, a theory begins, in my view with a hypothesis or conjecture which is different to a postulate or assumption. A theory is more than a conjecture though.
The Wikipedia entry for – Theory says:
A theory is a contemplative and rational type of abstract or generalizing thinking, or the results of such thinking … Depending on the context, the results might, for example, include generalized explanations of how nature works …
In modern science, the term “theory” refers to scientific theories, a well-confirmed type of explanation of nature, made in a way consistent with scientific method, and fulfilling the criteria required by modern science. Such theories are described in such a way that scientific tests should be able to provide empirical support for, or empirically contradict (“falsify”) it. Scientific theories are the most reliable, rigorous, and comprehensive form of scientific knowledge … Scientific theories are distinguished from hypotheses, which are individual empirically testable conjectures, and from scientific laws, which are descriptive accounts of how nature behaves under certain conditions.
So we get an idea that theory is about abstraction intending to explain characteristics of reality.
However, while a conjecture is a proposition we consider, in advance, will help us understand and explain reality, they transcend to theoretical status when they are confronted with that reality using data and empirical techniques.
We start with a conjecture or hypothesis and then interrogate that supposition with the data.
I don’t want to get into the eternal debates in the philosophy of science about falsification and the like. Suffice to say, I am not a Popperian (falsification) and lean more to the ideas of Imre Lakatos (but not his anti-Marxism British phase!) and even Paul Feyerabend.
A body of thought often doesn’t subject itself to empirical scrutiny because it is inherently untestable – it represents a sort of core faith – the ‘hard core’ of the research program in the Lakatosian sense.
Surrounding that ‘hard core’ is a ‘protective belt of auxilliary hypotheses’, which taken together constitute a falsifiable theory.
But the practice of refutation (empirical scrutiny) might conclude that the theory (the ‘hard core’ and the ‘auxilliary hypothesis’ together) produce false empirical predictions (see below) but then we have a dilemma.
Does the empirical prediction failure lead to a rejection of the ‘theory’ or just an alteration of the ‘auxilliary hypotheses’?
Scientific communities are very reluctant to jettison their ‘hard core’ when they can simply alter the ‘protective belt’.
But, ultimately, too many empirical failures, spawns counter paradigms, which eventually supplant the dominant program.
And so it goes.
But all those rabbit holes aside, what is obvious is that even though research and theorising in the social sciences must involve empirical confrontation, it is quite different to the natural or physical sciences because we study humans in uncontrollable situations.
That means the way we proceed and the types of conclusions we draw are quite different. There are no ‘laws’ in economics as there are in physics, for example.
This has been a debate within economics for decades.
The mainstream like to think they are doing ‘science’ akin to the way a physicist does science. This gives them a sense of authority, which they use to parade over the masses, and tell them that certain propositions are non-contestable.
They even call some of their conjectures – ‘Laws’ – for example, the so-called Law of Diminishing Returns, which is a pillar in the propositions that mass unemployment is the result of excessive real wages and, hence, cuts are required to cure it.
To hide the ideology, they invented a ‘Law’.
While gravity is a law founded on measurable physical relationships, this mainstream economics ‘Law’ is just an assertion, which, because of its conceptualisation and abstraction (marginal analysis), bears little relationship to real world practices and is inherently ‘untestable’ as a conjecture.
The real world doesn’t look as though the ‘Law’ could be applicable in most situations (certainly outside of agricultural applications).
So we understand that economics is a social science and is contestable as a result.
A theory is thus a logical ordering of linked conjectures about the real world.
A theory will typically have several components.
1. Definitions and concepts – which describe or delineate key components that make up the theory – so concepts such as national income, unemployment, trade etc
These concepts will be defined in particular ways and often, in economics, related to in an accounting manner.
So total output (GDP) is equal to the sum of the expenditure flow on final goods and services in some particular period.
There are a number of definitions and concepts required to make that statement operational.
2. Functional or behavioural relations – which take the important variables (concepts) and write out how they interact.
What drives what? And to what effect?
This stage of the theory requires us to make assumptions as a way of adding some control to the analysis and to develop conjectures (hypotheses).
3. Predictions and shocks – we then ‘shock’ the behavioral relationships that make up our ‘theory’ by changing values of variables or the parameters linking them in the behavioural specification and make predictions as to what happens to some or all of the variables (the ‘system’).
That is what I think of as a process of theorising.
A theory transcends a conjecture when it has empirical congruency.
So if we say that if national income rises by a dollar, household consumption will rise but not by a dollar – that is a conjecture.
If it turns out that the conjecture is what happens in the real world, then we have a theory that is congruent.
Predictive accuracy – congruency
When we undertake empirical research in economics using the techniques derived from mathematical statistics and econometrics we are not searching for truth but congruency with the data.
Congruency means that an explanation is currently the tentatively most adequate in terms of helping us understand the dynamics of the data of the real world.
There is a lot of misconception about that.
A congruent body of ideas should also have some predictive confidence, although cannot withstand sudden regime shifts the push the data in unpredictable ways.
I wrote about those themes some years ago and this 2007 Working Paper (a revised version of a paper I wrote in May 1995) – Econometrics, Realism and Policy in Post Keynesian Economics – covers some of the discussion. This paper was subsequently published in a modified form but this version is free to all.
The point of the paper was to oppose the view by so-called Post Keynesian realists who eschewed any serious empirical work in economics on the grounds of fundamental uncertainty.
I argued that econometric research and practice is a necessary part of a paradigm which ultimately emphasises policy intervention.
I argued that, without negating the existence of fundamental uncertainty, the demands of policy can be accommodated and the utility of econometrics justified within a satisficing environment where rules of thumb and habits create a stable context for decision making.
If you take the fundamental uncertainty line to its logical conclusion you end up with a a nihilism, which would appear to be at odds with an interventionist policy emphasis.
Following the great George Schackle, it is true that “we simply do not know” about the future but we make guesses about it and try to calibrate those guesses based upon what we do know – humans behave habitually, hunt in packs, etc.
Fundamental uncertainty only prevents us from ever discovering, knowingly, the truth and therefore the best solution.
As a result, people and institutions thus search for adequate solutions to their problems. In the institutionalist literature this has been called satisficing. When we discuss satisficing the role of rules of thumb and habits emerges.
It is not the existence of uncertainty which is important, but rather the ways in which people in the economy deal with its existence.
So when I talk about predictive accuracy, I’m not talking about knowing the future or establishing truth. Rather, I’m talking about congruency.
And I think it is without doubt that economists who see the world through an MMT lens have been much more successful in making statements about future events, particularly following major policy interventions, than our mainstream colleagues.
The latter have what can be called an appalling record of congruency.
There was a great article in The Week last month (August 14, 2018) by Ryan Cooper – The biggest policy mistake of the last decade – which took his readers through a sequence of forecasting blunders by significant mainstream economists after the GFC.
He notes the predictions of Alberto Alesina and Silvia Ardagna, Carmen Reinhart and Kenneth Rogoff, apart from having “major conceptual problems”, “got everything wrong” about the real world.
He also cites “a bunch of conservative financial and economics luminaries, including Michael Boskin, John Cogan, Niall Ferguson, Kevin Hassett, Douglas Holtz-Eakin, Bill Kristol, and John Taylor” who were also raising alarm bells about inflation following the fiscal stimulus in the US:
Yet it’s been six to eight years since their arguments and there’s hardly been a glimmer of the kind of inflation they warned about.
In fact, not only has there been no hyperinflation, inflation has consistently come in under the Fed’s supposed target value of 2 percent.
He lists other major blunders by the mainstream doomsayers.
I discussed some of these issues in this blog post – MMT predicts well – Groupthink in action (February 6, 2017).
Abstraction is an important part of theorising. We cannot simply have a theory of everything (easily!). So we simplify.
In itself simplification is not a reason to criticise a theoretical structure. So attacking neoclassical ‘theory’ because involves highly abstracted assumptions to be tractable is not the reason we reject it.
We reject it, for example, because the predictions it makes about human behaviour, which are based on their highly simplistic set of axioms about human behaviour, simply do not explain reality.
Psychologists just laugh at the promotion of Homo Economicus (the basis of neoclassical theory about the individual) by mainstream economists.
We should also understand that in macroeconomics there are many abstract concepts.
If you think about it (and understand it), the entire body of macroeconomics is an abstraction.
There is no such thing as a “price level” in the aggregate, or a household. The aggregates are the sum of the parts.
So that sort of abstraction is not the reason to criticise a body of work.
However, if those abstractions, taken together, lead to predictions that are just ridiculous, then one can reject the ‘theory’.
For example, New Keynesian macroeconomics did not even have the banking sector included before the GFC. It could never explain the reality that unfolded as a consequence.
Now why is there a T in MMT?
At the outset, we have to be careful not to conflate theory with new. A body of work can be considered to be a theory even if in putting together that body of work there was no new theorising involved.
In the case of MMT there is a substantial body of ‘theory’ that we inherited from past schools of thought and integrated into a MMT framework.
The standing on the shoulders of giants syndrome!
Now, lets start thinking about theory in the MMT context.
Here is a simple example.
An oft-stated claim is that MMT is about accounting relationships. Well in part it certainly is. It is part of the stock-flow tradition in heterodox economics that ensures consistency between flows of things and the stocks they flow into, period to period.
I wrote about that in this blog post – Stock-flow consistent macro models (September 8, 2009).
Those who make that spurious claim about MMT often use the sectoral balances framework to make their point.
They note that the basic sectoral balances relationship, which is a core part of the way an MMT economist analyses the world, is, at heart an accounting truism that has to be true because it is derived from a larger accounting framework – the nation’s National Accounts.
That is true so far.
While I know there is a debate in accounting about the theory of accounting, we will accept, here, that an accounting truism is one that has to be true (add up in this case) by the way we define it. It is not opinion or conjecture – it just has to be.
So the statement: the Government deficit (surplus) equals the Non-government surplus (deficit) dollar-for-dollar is such a truism.
It must be true.
By introducing this truism into the economic debate, MMT proponents have been able to alter the way people think about government deficits and surpluses and that sparks interesting new insights and options.
So when a mainstream economist or a politician that is mouthing that same economist rails against deficits and says surpluses represent national saving and putting money into the piggy bank for the future, anyone who understands MMT knows automatically that surpluses destroy non-government financial wealth (through the fiscal squeeze).
They further know that a currency-issuing government doesn’t ‘save’ in the sense that a household saves by postponing consumption to generate higher consumption possibilities in the future.
A financially-constrained household has to save to achieve that goal.
A currency-issuing government can spend whenever it wants as long as there are things for sale in its own currency, irrespective of what it did last period.
A fiscal surplus provides such a government with no extra capacity to spend in the future.
So a simple truism has a powerful effect if understood correctly.
But we can go much further than that.
When we write out the sectoral balances equation as:
[(S – I) – CAD] = (G – T)
We know the term on the left-hand side [(S – I) – CAD] is the non-government sector financial balance and is of equal and opposite sign to the government financial balance.
So if (G – T) > 0 (a fiscal deficit) then it has to be true that [(S – I) – CAD] > 0 (a non-government surplus).
Further, if (G – T) increases (that is, a larger fiscal deficit), then the non-government surplus has to increase.
These propositions are true by definition.
Further, if there is a current account deficit (CAD < 0), and the private domestic sector start to save overall (S – I > 0) then the government deficit has to become larger and if the government tries to resist that fall in non-government spending, then national income will fall and the fiscal deficit will become even larger still.
This statement goes beyond the inherent sectoral balances accounting.
Focus on the part of that statement – “national income will fall”.
The interesting question is what drives the movements in the components of the sectoral balances such that they always obey the accounting truism?
That is where we need theory!
Remember that the derivation of the sectoral balances comes out of the basic income-expenditure framework that is a core part of MMT (inherited from earlier Keynesian theory).
We know that expenditure drives income through output creation.
In the national accounts, the sources of expenditure are:
1. Household consumption expenditure (C)
2. Investment expenditure (I)
3. Government expenditure (G)
4. Exports (X) less Imports (M) – or Net Exports (X – M)
We can clearly write out accounting relationships (and identities) about these variables such as
(1) GDP = C + I + G + (X – M)
which says that total national income (GDP) is the sum of total final consumption spending (C), total private investment (I), total government spending (G) and net exports (X – M).
Expression (1) tells us that total income in the economy per period will be exactly equal to total spending from all sources of expenditure.
But that begs the question, what determines (explains) these components (C, I, G, X, M)?
To answer that question, we need theories about the behaviour which determines how much and when a household or business firm will spend.
We need theories of government spending.
We need theories about the way trade operates to determine exports and imports.
We also know that when the government taxes total income, households are left with disposable income. That is just an accounting statement.
Then households make choices on how to use that disposable income – to consume and by construction (as a residual) to save. For the individual household total saving becomes a behavioural choice. We need a theory to explain that.
In the most simple MMT expenditure model (which we inherited) we say that
1. C is a function of disposable income. That seems like a very simple statement. But it is a conjecture that can become theory if it stacks up against the data.
More advanced studies try to work out the nature of the relationship between C and (Y – T) and a number of competing theories are out there (Permanent Income Hypothesis, Relative Income Hypothesis, Life Cycle Hypothesis and more).
We might try to work out how wealth (a stock) impacts on consumption expenditure (a flow).
2. I is a function of the cost of funds (interest rate), expected revenue (proxied by some function of income) and other non-linear type of impacts (irreversibility etc).
And so it goes. I won’t rehearse all the theories for G, X and M (only M is considered in the simple model to be a function of income and real exchange rates).
So while the sectoral balances equation as an accounting identity provides some interesting insights in its own right, the really interesting times begin when we place the equation in a dynamic setting to analysis what happens if.
The ‘if’ being a change in the one or more component and the way those changes evolve over time by reverberating through the theoretical linkages between the components.
And by theorising we understand how national income changes bring the economy into a state that we account for.
Thus, if private domestic sector households decide to cut consumption expenditure, what happens next?
We need our theory to trace the consequences through!
Firms notice inventory build up. They don’t produce to store goods and servies. They produce to sell.
So they start to cut production – first by cutting working hours (shifts, intensity etc) then, if the expenditure decline cements itself, they start laying of labour.
Incomes fall and this reverberates into further reductions in expenditure.
Soon enough the firms form a view that expected future revenue from sales is lower than previously expected and so they cut investment plans.
And this creates further expenditure shocks and so on.
Our theory tells us that the national income changes that follow these expenditure adjustments impact on:
1. C – it falls.
2. S (household saving) – it falls because it is linked to income.
3. I – it falls.
4. T (taxes) – they fall because they are linked to economic activity (income generation).
5. M (imports) – they fall because they are linked to economic activity (income generation).
6. G – it is likely to rise because welfare payments through income support systems rise.
And so you can see that the national income changes drive shifts in the financial balances for each sector.
At all times in this adjustment path, the accounting relationship we call the sectoral balances equation is satisfied.
Eventually, when the economy reaches a new equilibrium (an abstraction), which just means the national income changes from the expenditure shock finally become zero, we reach a new set of sectoral balances.
But we needed theory to explain how we got from one steady state to the next one following a shock.
Of course, in the real world, there are shocks occuring continuously, and it is hard to isolate them.
But that doesn’t alter the point.
I could talk about other examples of how ‘theory’ is a core part of what we call MMT.
I discussed the ridiculous proposition that MMT has no inflation theory in this blog post – I wonder what the hell I have been writing all these years (February 12, 2013).
The same character who made the original assertions that MMT had ignored inflation is still making the claims – he is just a serial nuisance.
But as I indicated yesterday, MMT has an elaborate theory of inflation embedded within it.
Part D of our textbook, the front page of that part shown here, contains 3 chapters which develop that theory in contradistinction to the mainstream theory.
We inherited some of the ‘inflation’ theory that is now part of the core MMT body of work. But some of it is new.
I can claim some credit there personally as key propositions from my PhD (buffer stocks etc) are now part of that inflation approach and represent a ‘new’ way of thinking about the relationship between unemployment and inflation.
MMT replaces the mainstream Phillips curve with an employment buffer stock.
See the linked blog posts I provided at the top of this blog post for more discussion about that.
There are many other examples of how MMT combines accounting, description and theory.
But I think I have given you the idea.
MMT is much more than a descriptive lens.
It is a coherent body of propositions that allow us to understand how the monetary system works. Understanding comes with explanation.
Explanation requires, in part, theory.
MMT is appropriately named. There is no misnomer.
That is enough for today!
(c) Copyright 2018 William Mitchell. All Rights Reserved.