An Actually Accurate Criticism of MMT

MMT is Narrow in Scope

I've been discussing on twitter of late, how I would like to move past the "MMT" label, in discussing economics.  The problem is not that MMT is wrong or flawed, but that it is only concerned with a specific set of issues related to public finance.  Furthermore, it relies on a somewhat simplified "money story", which I believe is useful for understanding how money works, but which is ill suited as a more general framework to understand political relations and economic development.

The problem with the MMT money story, is that it starts with taxes, and not the issue of "what is property?".  I think this issue of what is property is more important politically and socially, but talking about money in terms of taxes is the quickest way to get an idea of what money is, and what function it serves.  So it is not that I find the MMT money story flawed, on the contrary, only that it is limited by its focus on money.  If your goal is explaining money, then this "tax-driven" story may be the best description, but learning MMT quickly helps one realize that the intricacies of accounting are not what limits or empowers societies and the members of said society.  Here are some points often emphasized in MMT in its tax based money story:

1. Money is a tax credit (including bonds)
2. Unemployment is a bid to buy money with labor.
3. Taxes create unemployment(intentionally)
4. Therefore, taxes drive money

The first three statements are axioms, and the fourth is the conclusion.  These four statements should be enough for most anyone so inclined to derive the rest of the concepts and framework used in MMT analysis.  Importantly, characterizing money as a tax credit means that sovereign money issuers do not have any debts in their own currency.  It is not that the sovereign seeks financing, but rather that it offers different financial instruments for savers.  Both the treasury and federal reserve(in the U.S.) can create money, although the treasury is constrained to creating less liquid bonds, and the federal reserve is obligated by its mandate to offer an exchange rate between currency and bonds, namely the rate of interest.

A refusal to monetize bonds at any price, is in effect, setting the interest rate to infinity.  An infinite interest rate is outside the scope of fed authority, because that would be a tax.  So central banks are obligated to convert between bonds and currency at some finite rate of interest.  While a central bank may refuse to redeem a mature bond themselves, the corresponding treasury can always issue a forward dated bond, which the bankers would then be obligated to redeem to maintain the rate of interest.

Many people accuse MMT of not having a theory of inflation.  But I think this accusation shows how shortsided most people are when they think about economic theory.  If MMT did not have a theory of inflation, that would likely be strength, because pricing and inflation are complex phenomenon involving many aspects from resource usage to political structure, to the distribution of consumption, to social power or even cultural phenomenon.

But in fact, MMT has a very clear and specific theory of inflation, once you get past the first layers of "mantras", which seem to define modern monetary theory.  This theory is summarized clearly in Mosler's book Seven Deadly Innocent Frauds of Economic Policy, where he quotes his earlier book Soft Currency Economics.

 The price level is a function of prices paid by government when it spends, or collateral demanded when it lends.

So whereas some commentators have been inclined to compare MMT to the "fiscal theory of the price level", MMT is better characterized as "the fiscal BIDDING theory of the price level". This slight change in verbage makes all the difference.

 It would be a mistake to completely reduce MMT to this "fiscal bid" inflation theory, but it is there, and Mosler and Mitchell have both discussed it in depth, as well as being related to proposed JG programs.  In particular, Kelton uses a slightly different language when discussing inflation, and that is that inflation may be evidence of a deficit that is too large.  In her writing, Kelton prefers to take a more conventional approach to inflation analysis, in that many inflation episodes may be attributable to "demand-pull" or "cost-push".   While I can't speak to her motivations, I suspect she chooses this language deliberately for good reasons.  The fiscal bid theory of inflation, faces a couple challenges before it can be presented as proven.  First, it may be politically impractical, in that constraining government's bids limits its ability achieve fiscal priorities.  In 7 deadly innocent frauds, Mosler's example of price anchoring has government spending dropping suddenly to zero from one year to the next:

Let me give you an extreme example of how this works: Suppose the government said it wasn’t going to pay a penny more for anything this year than it paid last year, and was going to leave taxes as they are in any case. And then suppose this year all prices went up by more than that. In that case, with its policy of not paying a penny more for anything, government would decide that spending would go from last year’s $3.5 trillion to 0. That would leave the private sector trillions of dollars short of the funds it needs to pay the taxes.

In addition to the practical importance of bidding processes for fiscal prioritization, there is another glaring issue with MMT's unique inflation theory.   Fiscal price anchoring has never really been implemented at scale, so there is a certain aspect of speculation in trying to guess how this program would play out.  In a theory sense it is the perfect inflation control, allowing you to always spend without limit at a specific bid, so long as someone is willing to fulfill that bid. But in practice, it is hard to assess whether it would create dramatic cyclical instabilities or disrupt normal government operations.

So MMT definitely has a theory of inflation, which is specific(maybe too specific), but has not been tried, and in the worst case results in an unhelpful tautology: "We only got inflation because government increased its bids for goods and services", when one might reasonably conclude the government necessarily and prudently increased bids to maintain critical operations.  Later I will discuss the Job Guarantee, which has the important property of being non-critical but hopefully beneficial labor, and thus sidesteps these bidding concerns for essential operations.

In the interest of completeness, there is a third kind of inflation which I believe many authors fail to recognize is even possible: and that is "price drift".  So there is not too little supply, nor too much demand, but over time prices can just "drift" off of their historical points, without any specific driver.  Any specific price change is too small to be linked to a cost or demand phenomenon, but over time such drift can have a bias.  In an extreme case, this drift could become quite rapid, which is what I believe inflation "expectations" really are, simply price drift that has gained momentum.  Price drift including rapid price drift, is not a problem in itself, except it may make it difficult for everyone to keep up, or present some sudden issues of crashing.  Importantly, price drift can lead to equivalent balance sheet positions.

Consider an economy with 2 people and 2 goods.  We will present two scenarios, describing a very crude "balance sheet" and accompanying "price structure" for each.

Scenario A

Bill: $30, Ted: $70.

Apples: $2, Oranges: $1.

Scenario B

Bill: $60, Ted: $140.

Apples: $4, Oranges: $2.

Both the above scenarios are equivalent.  Importantly, both the balance sheet positions and the relative price structure must be isomorphic(a specific mathematical property between systems with equivalent structure but different labels or scales).  But if that is the case, it does not make a difference whether you live in scenario A, or scenario B.  If a world were to move from scenario A to scenario B, that would mean inflation happened, but it would also be a completely benign inflation, in other words, price drift.  In practice, some parties might be winners and some losers, from the asymmetries of such a transition to a higher price level, but such temporary advantages could be limited if the prices stabilized fairly.

So this third kind of inflation, price drift, should be important to any meaningful discussion.  Additionally, discussions of demand push and cost pull are incomplete if you neglect to talk about price setting power.  And this is where MMT completely copies its answers from previous scholarship, this issue of price setting power is a central theme behind post-Keynesian scholarship and some heterodox and even orthodox economic writers.

MMT's Job Guarantee is a Policy Model

So now we have discussed the fiscal bidding theory of the price level, we come to the proposed mmt solution: price anchoring.  The premise of price anchoring is to have a stable government bid for goods and services.  This means in particular that if market prices rise, government budgets automatically contract, or if market prices fall, then government budgets may need to expand, to prevent unemployment.

The Job Guarantee is the example of an idealized and universal price anchoring scheme.  It pays the minimum wage, or rather, the minimum wage is just whatever the job guarantee pays.  In this sense the Job guarantee is a theory construct designed to achieve full employment and price stability, but it does use a different definition of what full employment means.

Some accuse MMT of not having any models, or even intentionally being vague to be impossible to disprove, but that is precisely the purpose of MMT's Job Guarantee proposals, to provide a clear example of how "price anchoring" would work in practice, and give people a specific proposal to analyze. While many do literally want a Job Guarantee, its relevance toward theory is how it serves as an "employed buffer stock", demonstrates a fiscal "employer of last resort" program, and captures the essence of "price anchoring".

In the face of inflation, the Job Guarantee wage should not be raised, or even possibly lowered, to stabilize the macro economic system more rapidly, without the associated costs of "involuntary" unemployment.  And it is here where one might take the most issue with MMT.  MMTers use a very different definition of unemployment than the rest of the world, in that they nearly always specify "involuntary unemployment".  This seems to be suggested as people willing to accept a minimum wage, given amenable working conditions and terms.

This focus on involuntary unemployment may be reminiscent of another online group, that of the "involuntary celibates". But unlike the infamous internet incels, MMT's "workcels", with ample free time to complain on public forums, have a much less ambitious ask, to be employed at the minimum wage in activities they deem to be socially beneficial.

Many of both groups see themselves as victims of a biased and unfair system.  Incels may have difficulty finding sexual partners, but the inemps or workcels may struggle finding a capitalist willing to engage in wage/work exchange.

MMTers thus create a unique definition of unemployment, rather than merely someone searching for work, the definition of involuntarily unemployed are willing to work for the lowest wage ask, albeit what they can bring to the table is incidental, and the conditions of employment are critical.  If a job is too far away, too demanding, or requires unreasonable hours, then it is not the worker who has failed, but the employer who is unreasonable.

It is not surprising that individuals find the conditions of modern work frustrating.  Some people are greatly accomodated, while others are exploited and treated poorly.  The reasons for this disparity are not always obvious, and not as simple as what those supporting a free market fantasy would suggest.  Differences in pay are not always differences in worker productivity or value.  There are many other incidental issues as well.

Moving Forward

I wrote this critique, not only because the critiques of MMT are so frustratingly nonsense, but mainly so that moving forward, in an attempt to model MMT principles, I can tell critics that I have seriously considered the weaknesses of this theory.  At the very least, it would be useful to distinguish cohorts of unemployed people by their wage ask.  Capitalism is supposed to give us creative destruction, which can render even very skilled and valuable workers obsolete at a moment's notice.

Is the solution to have government provide jobs for everyone, albeit for a very modest wage?  I cannot say for certain, and I don't even think this is the most interesting part of MMT, only the simplest way to present its premise.  Hopefully I can get you some good solid models of price anchoring that are not a job guarantee, and thus demonstrate that MMT ideas may have relevance outside.

In any case, the minimum wage has been one of the least politically controversial policies, but most controversial in terms of economic theory.  There is a simple explanation that meets in the middle.  Because labor time is perishable and local, it is not surprising that wage asks may depart from the equilibrium.  A minimum wage sets a standard for both wages and worker productivity, to prevent a "slippery slope" effect from systematically devaluing wages, and an "outlier effect", where a small number of people might otherwise work for less than the prevailing "starting wage offer".

I prefer the term "starting wage offer", or "basic work wage" to the phrase "unskilled labor".  If computer science has taught me anything, it is that very basic skills are very impressive, being able to walk, talk, multi-task etc.  These basic skills are what we do well compared to machines, which completely changes what one might see as skilled or unskilled.  Any underpowered arm processor with a stockfish engine can beat magnus carlsen, but even skilled robots will struggle to perform a hand shake.

MMT and Mathematics

If MMT suffers from a lack of mathematical development, that is mostly because the pool of MMT economists is much smaller, and many MMTers are part time or casual enthusiasts, and not professional economists with time and resources.  Beyond that, there is something to be said, that mathematical modeling on the subject of macroeconomics presents particular difficulties.  At this level, all models must simplify, but most are pure fantasy.  There is a difference between simplifying assumptions and fantastical ones, and while "simulations" make great mathematical models, very neat and nice to analyze, they make very poor historical ones.

Understanding historical phenomenon is not about reducing political and financial realities to a semi-realistic looking simulation.  There is a way to use modelling to formalize ideas and intuitions about how parts of the economy may interact, but I think to do so with the premise of realism is misguided.  Modelling should be about testing real world principles and ideas in a controlled laboratory, in order to demonstrate that our notion of a policy or program, even fails to be realistic in this controlled environment.  So modelling may be more useful for debunking reductive thinking, rather than promoting specific hypotheses for how the economy works.

And I think here is where economists frequently make a fatal type error in their analysis.  The quantities measured by accounting, accounting variables, are not direct measurements of the economy itself.  So prices, interest rates, and inflation, have no meaning in terms of first principles, and no specific interpretation.  These are accounting phenomenon, not economic ones.  And this is how we "see through the glass darkly".  We would like our accounting to represent something real and accurate about the real economy, and some of these phenomenon like inflation reveal there is a disconnect.  But to study inflation and unemployment like it were the system, and not a representation of the system, is wrong.  In particular, from this lense, the phillip's curve becomes a simple perverse condition where working more(lower unemployment), may make us poorer(increases inflation).  In that sense, it should prompt us to evaluate our productive priorities and the way we create employment, and not try to infer some universal principle, just because we suck at working in the margin.

Despite challenges and limitations, there is some work of a mathematical nature being done by MMTers and those sympathetic. Sam levey has developed mathematical models, and steve keen has even made a custom software package called Minsky for of modeling economic dynamical systems.  While both these efforts are significant in formalizing certain ideas, I suspect that most economists outside MMT would not really find any value there.  The most significant challenge of mathematical modelling is one of representation and scope: how are you going to represent these parts of the accounting and economic machine, and what do you hope to infer from this model?  As I stated earlier, serving as a laboratory to refute simplistic analytical thinking may be more productive than expecting a model to comport with or test reality.  In other words, the economy can be at least as complex as your model, but it is not limited by it.  In some cases the economy may be very simple but spontaneous.  People decide they don't like something, and they get rid of it.

The main reason people ask for mathematical models of MMT, is that they are too lazy to actually learn the ideas.  MMT has a theory of inflation, but that theory is untested.  MMT does not have a unified simple theory of interest rates, besides the notion that interest is regressive.  While mosler and many others, including myself, emphasize neofisherian ideas, it would be a mistake to say that all MMTers believe higher rates are always inflationary.

I want to reconcile on interest rates

I think that nominal rate increases can be used to communicate changes in the real terms of credit, that bond yields are acceptable so long as they are bounded, yet the people targeting so called "real rates" are delusional lunatics who favor capital to a pathological degree.  MMT has no simple or unified theory of how interest rates work, only the stipulation that real returns on public money are regressive and unnecessary under an ELR/JG framework.

Real rates on public money are regressive. Despite anything one might critique of MMT, I agree with this, and that is why it is so important to limit the returns on treasuries.