As the public scrutiny of the body of work we now refer to as Modern Monetary Theory (MMT) widens there is a lot of misinformation abroad that distorts or otherwise undermines what has been done to date. Most, but not all the misinformation or emphasis comes from those who attack our work. Their criticisms usually disclose an incomplete understanding of where MMT came from and what the core propositions and logic are. They stylise, usually using terms and constructs that are present in mainstream thinking, but inapplicable to an MMT way of thinking, and end up spitting out things like ‘printing money’ etc, which they think represents a devastating rejection of our work. As part of my own work, and I do this in liaison with Warren Mosler, I am interested in documenting the train of events that led to what we now call MMT. I love history and think it is very important in helping us understand things. So today I am continuing to examine archives to trace the provenance of key MMT concepts. And I am continuing to document the idea of a Job Guarantee, which is central to the MMT framework, despite many who claim to be MMTers thinking otherwise. I have noted in the recent press, claims that the origins of the buffer stock employment approach that became the Job Guarantee was the work of Hyman Minsky. Nothing could be further from the truth as you will see. It is important, in my view, to make the provenance very clear and that is what this blog post is about.
I have been providing factual historical information over the years to establish the provenance of the body of work.
Please read, for example, the following relatively recent blog posts:
1. Flattening the curve – the Phillips curve that is (April 7, 2020).
2. The historical beginning of the MMT team – from the archives (November 27, 2019).
In this vein, I was concerned to read in the Financial Times article (April 17, 2020), which reported an interview with Stephanie Kelton, that the work of Hyman Minsky was the source of the inclusion of a Job Guarantee in the development of Modern Monetary Theory (MMT).
Whether this was misreporting or not, the contention is false and amounts to historical revisionism.
As a general point, I provided an assessment of the relevance of Hyman Minsky’s work to the development of MMT in this blog post – Hyman Minsky was not a guiding light for MMT (November 9, 2016).
I argued that earlier in his career, Hyman Minsky advocated Abba Lerner’s functional finance.
But later in his career, Minsky seemed to shift and started to advocate what we now term to be ‘sound finance’ principles merging with a softer ‘deficit dove’ narrative.
None of his later positions are consistent with the core principles of MMT.
His 1986 book – Stabilizing an Unstable Economy (Yale University Press) – appeared to be a turning point. Minsky’s ‘sound finance’ path really emerged in that book.
Whereas he had previously considered government debt to be a stabilising force (to allow banks to diversify into risk-free assets etc) he changed his view in the 1980s when Ronald Reagan was President.
At this point, he increasingly argued that government debt was at risk of becoming non-credible in the face of non-government bond investors.
In his 1986 book, he wrote (pp.302, 304):
A government can run a deficit during a recession without suffering a deterioration of its creditworthiness if there is a tax and spending regime in place that would yield a favourable cash flow (a surplus) under reasonable and attainable circumstances …
Any deviation from a government budget that is balanced or in surplus must be understood as transitory- the war will be over, the resource development program will be finished, or income will be at the full employment level
His main message was that the fiscal outcome should be in balance or in surplus at full employment. However, that is not the MMT position at all.
He amplified that message in April 1991, when he produced a Working Paper 51 for the Levy Institute – Financial Crises: Systemic or Idiosyncratic – which was a paper presented at a Levy conference in that month.
The paper was about how the financial system could be ‘fixed’ in the face of increasing uncertainty from events such as the Savings and Loan crisis in the 1980s.
In the blog post cited above, I provide detailed analysis of his shift in thinking.
But this quote is representative.
In terms of the government sector, he wrote (p.28):
The government is no different than any other organization in that it needs revenues to validate its debts. This means that the government should have a normal conditions balanced budget, allowing for deficits in recessions and depressions and major wars.
This is not even remotely an MMT position.
But what of the claim that Minsky is the source of the Job Guarantee idea as it is expressed in the MMT literature?
Minsky on Full Employment
In 1965, Hyman Minsky contributed a chapter – The Role of Employment Policy – in the volume edited by Margaret S. Gordon – Poverty in America. The download link is provided by the Minsky Archive, maintained at the Bard College. It is a great resource.
This chapter is cited as being the source of Minksy’s advocacy of a ‘job guarantee’.
In that chapter, Minsky correctly noted that a “war against poverty” has to include “a program of job creation” and that required a nation to “sustain tight labor markets”, which, in turn required:
… bolder, more imaginative, and more consistent us of expansionary monetary and fiscal policy to create jobs than we have witnessed to date.
The problem that he recognised was to maintain “tight full employment” without “undue inflation”.
The paper was about how to achieve that goal.
His concept of “tight full employment” required that “at going wages”, firms “would prefer to employ more workers than they in fact do”.
In other words, vacancies outstrip the unemployed, which used be the way we conceived of full employment.
Such a state leads to dynamic efficiencies: higher wages, upgrading of jobs from low wage to higher wages, higher productivity, higher participation and lower unemployment.
Arthur Okun talked about this in his ‘upgrading hypothesis’.
Minsky, however, argued that pursuing “tight full employment” would “lead to a cost push inflation” arising from the wage upgrading effects.
However, that inflation would be a once-off adjustment to the new wage relativities.
His paper then focused on the “barriers” to achieving “tight full employment” and it is hear that he talked about the ‘gold standard’ and “our balance of payments problem”. This related to the idea that in a fixed exchange rate system (as it was then), monetary policy had to target the exchange rate.
He concluded that:
Thus, the only really available devises for expanding aggregate demand are fiscal.
And the problem then, in a Phillips curve world, was ensuring the fiscal expansion didn’t create inflationary bottlenecks – in other words, it had to be tailored to “have the largest primary and secondary impact upon the present poor.”
He criticised past spending initiatives as being “biased against the poor”. And he thought it was accurate to say that “the cross that the American poor bear is made of gold” – in reference to the balance of payments constraint which forced the government to eschew tight labour markets.
He clearly thought the US should abandon the gold standard but understood that would not happen at that time.
After rejecting the traditional “public works” approach to public sector job creation, he proposed several policy interventions to target the poor (such as, training, relocation, etc).
Among these interventions, Minsky wrote the following:
Work should be made available to all who want work at the national minimum wage. This would be a wage support law, analogous to the price supports for agricultural products. It would replace the minimum wage law; for, if work is available to all at the minimum wage, no labor will be available to private employers at a wage lower than this minimum … To qualify for employment at these terms, all that would be necessary would be to register at the local public employment office.
On the face of it, this proposal has similarities with the Job Guarantee.
But Minsky’s motivation was quite different – as a war on poverty. There was no mention of the scheme being an ‘automatic stabiliser’ or a redistributive mechanism to shift workers from inflating to the fixed price sector.
Indeed, Minsky was all about creating conditions where “low wages in the private sector would be pushed up, hopefully more quickly than high wages” once the scheme was introduced.
But even if Minsky had elaborated the full concept of what we call the Job Guarantee at that time, it would still not make the claim valid.
The provenance of the Job Guarantee concept in MMT
The FT article brought up the question of the Job Guarantee, which goes to the heart of the provenance of the concept within MMT and prompted me to write this response.
This is the relevant section (the interviewer writing in the first person):
I have always wondered why Kelton ties modern monetary theory explicitly to the policy of a federal jobs guarantee — a minimum pay cheque, for anyone who wants one. “It’s all in Minsky,” she says. A job guarantee is an “automatic stabiliser”, she explains. It stabilises growth by pushing money into the economy during a downturn in the most straightforward way: as firms cut staff, people still have a salary to spend.
Let us be clear – the contention that the concept of the Job Guarantee, a core MMT construct, is “all in Minsky” is false.
I will not attempt to impute motivation for the comments reported in the Financial Times here. Perhaps, there is some misquoting going on. Who knows?
But anyone familiar with the beginnings of MMT would not impute Minsky as a valuable source of motivation.
What I can say, with the authority of two (Warren Mosler and myself) is that the concept of the Job Guarantee as it entered the body of work which we now call Modern Monetary Theory (MMT) had nothing to do with the work of Hyman Minsky.
Minsky was in no way an intellectual source of the ideas that have become known as the Job Guarantee.
To say otherwise is to reinvent history and deny the basic provenance of the concept within the MMT framework.
And clearly, Warren and I are interested in preserving the provenance of the foundations.
To be absolutely clear, Warren and I have discussed this issue at length.
I have documented in the past the evolution of my own thinking that led to the proposition I called the Buffer Stock Employment (BSE) approach to full employment and price stability.
I won’t repeat that here but I recently discussed that, again, in some detail, in this blog post – Flattening the curve – the Phillips curve that is (April 7, 2020).
What I didn’t say then is that at the time, even up to the point that I met Warren and we more or less agreed that the BSE and his Employer of Last Resort (ELR) approach were the same, is that I had never really read any of Minsky’s work. That includes the 1965 article or any later work where he elaborated his poverty approach.
I had certainly not read his work on employment but had a familiarity with his ideas on financial instability, which have nothing to do with the topic in focus here.
I knew Warren also had come up with the ELR idea, independently of any reading of Minsky. But to ensure that recollection was accurate, I also sought specific clarification from him before I wrote this response.
The following interchange in the last 24 hours between us is relevant:
Q (me): When you were coming up with the original ELR idea, had you read Hyman Minsky’s 1965 article on full employment?
A (Warren): No, never even heard of him until somewhat later Randy brought his name up and said he’d set up a meeting. I asked Randy if he thought Minsky would take it in and he wasn’t all that optimistic.
Q (me): Was Minsky the motivation for your thinking at any point in those days?
A (Warren): No. The few parts I was directed to showed some possibility of an understanding of monetary operations, but others showed the opposite.
In this blog post – The historical beginning of the MMT team – from the archives (November 27, 2019) – I traced the beginnings of my relationship with Warren Mosler as contributors to the – Post Keynesian Theory E-mail Discussion List – which was an early Internet list server that began life in July 1994.
You can read the snippets from the archives in that blog post which helps to establish the provenance of the concept.
I used the term BSE to describe the ‘buffer stock employment’ model that I had developed in my undergraduate years in 1978. Warren introduced his Employer of Last Resort (ELR) model to the PKT list around then as well.
At this time, during the first two months of 1997, it became clear that the two conceptions were of the same thing and we started referring to the BSE/ELR approach.
In January 1997, I gave a ‘seminar’ to the discussion list (this was a process where one person would submit a paper and it would be discussed by the list for the next month).
In February 1997, it was Warren’s turn to give his seminar.
During those two months, there was a lot of push-back to the ideas from the list members including outright hostility from a few.
On Thursday, February 13, 1997, Warren Mosler wrote to the list and among other things said:
I wrote ‘Soft Currency Economics’ several years ago to outline why sovereign debt in local currency is always money good. It goes into the essence of why myself and several other hedge fund managers felt secure in purchasing Italian government lira debt at libor + 1.50% when most others were in fear of default. I had never heard the term ‘endogenous money;’ did not know the names Moore, Wray, Goodhart, and Kaldor; and knew nothing of Post Keynesian monetary economics.
Which reinforces the point that he was operating outside of the economics literature.
It was also clear that the majority of Post Keynesian economists (and the list was comprised of all the leading members of that school of thought, however, one wants to define it), were not particularly supportive of the BSE/ELR approach and failed to see the way in which a buffer stock would work.
After extensive criticism was aired, on Sunday, February 23, 1997, I wrote a detailed comment to support Warren where I reiterated the way in which the BSE model would work and noted among other things:
I cannot speak for Warren but my BSE model (which is equivalent in mechanics to ELR) certainly does see these jobs as regular day to day jobs providing dignity, security, and an attachment to the LF that i might currently have.
We were focusing on the place of the BSE/ELR in the monetary system.
Warren wrote later that day (February 23, 1997) in response:
This is exactly the type of dialogue all would be engaging in if they understood modern money and desired 0 unemployment and price stability. The political process would come up with a program and then modify it over time.
Our partnership was forming.
Paul Davidson, who was editor of the Journal of Post Keynesian Economics, and a lifelong supporter of Keynes’ own work, offered his summary of the main themes of the ‘seminar’ on February 23, 1997.
In part, he wrote:
… the ELR or what Bill calls the buffer stock of labor proposal. Here I think Bill and Warren have shown the best of the argument.
As an aside, to demonstrate the cognitive dissonance that was going on, even within this community of progressives, one member of PKT wrote (February 20, 1997):
… Mosler, Mitchell, Davidson and Wray are living somewhere else.
On February 24, 1997, Warren Mosler responded to a largely favourable rendition of his seminar paper with this:
… The traditional approach also works, but it takes an understanding of Bill Mitchell’s or Basil Moore’s last papers, for example to explain it. This is complicated enough so the man in the street votes to balance the budget. Including the most intelligent men in the street. The few econs against balancing the budget can’t get through with current logic.
On March 1, 1997, Randy Wray made his first intervention into the debate and wrote a long piece in support of the ELR approach.
He made the point at the outset:
I have long been interested in govt as employer of last resort because this was something advocated by Minsky (eg in his 1986 book). However, in Minsky it always seemed a bit utopian; there were a few gaps in the argument. When I saw Warren’s paper, however, he seemed to provide a way to fill the gaps
On March 7, 1997, the notion that BSE/ELR was intrinsically related to the broader Phillips curve debates in macroeconomics were made explicit.
I wrote (with edits for spelling and grammar):
In the development of my BSE model (which is similar to ELR) the important thing I am working on right now is the relationship between this option and the Phillips curve (PC).
Without the BSE in practice, there is a NAIRU – which is a reflection of a conflict between labour and capital over real income shares – that level of the unemployment rate, which temporarily renders the claims for real income compatible. It is not a harmonious or natural rate of unemployment rate – it might last, conceptually, for a day before some change to the bargaining relations sets the rate higher or lower for stable shares.
So there are a multitude of steady-state unemployment rates, each defined in terms of the past history of the economy (state dependence …) and each as unstable as the next.
The PC then is defined by a particular state-dependent path.
With a BSE policy, there is no unemployment so there can be no NAIRU in the broadest sense. There is only a private sector NAIRU …
The reality is with the BSE/ELR, wage-price conflict can still occur and so you can still get inflation. You cannot get macroeconomic unemployment though emerging from it.
But you can still get shifts between the private sector and the public sector … When the inflation spiral is eating into AD (as the govt does not accommodate the pressure), jobs are lost in the private sector but not overall. So we can still define some unemployment rate in terms of a current steady state (or compatible) bargaining relations. This unemployment rate is a private sector rate only. Significantly different inflation-unemployment rate (overall) dynamics emerge.
Warren responded on the same day and placed that conceptualisation within a micro-oriented monopoly:
Yes. I would add that fundamentally this stems from the government, as monopolist, setting (budgeting) q (spending and lending, and manipulating desired H(nfa) through monetary policy) and letting p (prices paid by govt.) be at market.
On March 9, 1997, Randy Wray reentered the conversation, saying:
… glad to see the elr seminar is carrying on; bill m. and warren have made some interesting contributions, but i want to print them so i can examine bill’s exposition in detail.
On March 16, 1997, Warren’s seminar, which had ‘raged’ for 6 weeks, was formally closed with a summary statement from Warren, which said, in part:
This will officially end the ‘Full Employment AND Price Stability’ (FEAPS) seminar. The focus was the elr proposal, which is not materially
different from Bill Mitchell’s BSE proposal … Perhaps what was added were reasons that an elr policy can allow the budget deficit to float while maintaining price stability as defined, much like various incomes policies have been designed to do. And that the key is for the Government to constrain the prices it pays, in this case for the elr wage, to maintain a stable value of its currency.
Once again, pitching the BSE/ELR in terms of the Phillips curve debate.
Once we (Warren, Randy and I) started working together and Warren had recruited others to the project (Mathew, Pavlina etc) we debated the terminology that we would use to describe the buffer stock approach.
As I have previously noted, this period coincided with the outbreak of the mad cow disease, principally in the UK, and so I thought it unwise to continue calling my approach the BSE model
I also didn’t like the terminology ELR for reasons I have explained elsewhere.
Around that time, I had read the paper of Wendell Gordon – Job Assurance — The Job Guarantee Revisited – which had just been published in the ”, Journal of Economic Issues – and I proposed we unify the BSE/ELR under the name Job Guarantee.
Gordon’s conception was not an MMT Job Guarantee, but I thought the terminology (JG) would help us advance.
Again, Minsky’s name or work, never came into the deliberations.
The point is that:
1. The concept of a Job Guarantee that is now core MMT was entered into the discussion at that time by Warren Mosler (ELR) and myself (BSE). This was the provenance of the concept within MMT.
2. Minsky was never mentioned. Only his former PhD student, Randy Wray, once exposed to the BSE/ELR ideas, noted some overlap between the BSE/ELR approach and Minsky’s own, earlier ideas. But that was well into the PKT debate about the concept and he acknowledged that the BSE/ELR approach that Warren and I were advocating went much further than Minsky’s ideas and tied the concept into core monetary logic – the MMT approach.
3. Anyone with knowledge of the history and the beginnings of the MMT work would not reasonably say that the reason that MMT considers a Job Guarantee to be an essential part of the body of work was due to anything that Hyman Minsky had written or said.
To try to argue that line is to explicitly deny (deliberately or through ignorance) the provenance of the concept.
It is important to render history as accurately as we can.
That is enough for today!
(c) Copyright 2020 William Mitchell. All Rights Reserved.