Today, our new Manga series - The Smith Family and their Adventures with Money -…
On October 2, 2019, I received an E-mail from Gregory Mankiw. It was sent to me, Randy Wray and Martin Watts and asked us some questions about our textbook – Macroeconomics – which had been published by leading textbook publisher Macmillan in March 2019. The book has been selling strongly with a third printing already in the pipeline and a second edition coming, hopefully, later next year. Macmillan also publish Greg Mankiw’s macroeconomics textbook, which has been the dominant teaching book in undergraduate programs. I will take you through the E-mail correspondence that followed because it puts in context what Greg Mankiw decided to do next. Instead of continuing the correspondence on academic terms, which was a reasonable expectation at the time, given the initial approach and our replies, he decided to submit a paper – A Skeptic’s Guide to Modern Monetary Theory (December 12, 2019) – to the American Economic Association meeting in early January, which purports to be a ‘guide’ (meaning in English – a framework to convey an appreciation of something) to Modern Monetary Theory (MMT). After his initial entreaty and our responses in good faith, Greg Mankiw clearly decided that engaging with us on the terms he initially set out was not going to be in his interests and thus took another tack, without any further consultation or reference to his initial contact with us. I wasn’t impressed with that strategy. I was less impressed with the ‘guide’ that emerged. It says very little about MMT. It demonstrates how hard it is for someone deeply locked into a dominant but failing paradigm to think outside the ‘box’ for a while and try to understand that the ideas of a new and emerging paradigm cannot be meaningfully reduced back into the conceptual framework of the failing paradigm that the contender is seeking to usurp. I guess his strategy is understandable – after all – our book is now a direct competitor for his textbook and offers a new approach that has much stronger empirical correspondence. In that context, it is in Greg Mankiw’s self interest to attack our book in any way he can. The problem is that attacks have to have some foundation to resonate. Greg Mankiw’s attack is so lateral that he would have been better to have remained silent. Sure, he is playing to the mainstream groupthink echo chamber. But the echoes will die eventually as more and more people realise the mainstream is in its last death throes. This is Part 1 of a two-part response to Greg Mankiw’s paper. In Part 1, we review the E-mail trail that started all this. In Part 2, I will discuss his response.
The E-mail trail
October 2, 2019
On Wed, Oct 2, 2019 at 4:35 PM Mankiw, N. Gregory … wrote:
Dear Professors Mitchell, Wray, and Watts,
I have been reading your book Macroeconomics. (Congratulations on writing it, by the way. I know how much work such projects are.)
To help me get my head around MMT, I have a question that I hope you can help me with. Here it is:
If you could design an experiment to distinguish MMT from conventional macro theory, what would it be?
What I am looking for is something along the lines of the following: If the government does X, conventional macro theory predicts Y, whereas MMT predicts Z. And Y and Z must be observably different.
The reason I ask is that I am wondering whether MMT offers different predictions or, instead, is a new conceptual framework for understanding conventional predictions. I know that running a macro experiment is likely infeasible in practice, but thinking about it might help clarify the theory for me.
Please let me know if there is an easy answer. And I apologize in advance for imposing on your time.
Which seemed to be a perfectly reasonable way to engage in a conversation. It was seeking input from us which suggested that there would be some conversation.
October 3, 2019
On Thursday, October 3, 2019, Randy Wray replied along these lines (not consulting either Martin nor myself prior to sending it):
Dear Greg: Thanks for opening a dialogue. Yes I think we can come up with a number of such hypothetical experiments.
I’ll just start it off with one.
Government increases spending such that G > T – what is normally called a spending deficit (a flow). The conventional view depends on the model – so let’s first take loanable funds and then ISLM.
a) The additional demand for funds pushes up interest rates and crowds out investment.
b) The additional spending flow raises money demand, pushes up interest rates, and crowds out investment.
I do know that you could come up with special cases where this did not happen but I’m keeping it simple.
The MMT view is that if G > T, bank reserves will have been net credited and hence the pressure on the overnight interest rate will be downward, all else equal.
Again, I am keeping this simple.
So the orthodox approaches predict upward pressure on rates, while the MMT approach predicts downward pressure.
We can then get into central bank reactions to these effects in both the conventional and the MMT approaches, and as well into actual procedures adopted by central banks and treasuries to effectuate spending and taxing operations. (ie there may be operations BEFORE as well as AFTER the sequences I outlined).
So that seemed straightforward and after reading that I assumed Randy was anticipating a conversation would follow in response to this ‘experiment’ which Greg Mankiw had requested.
October 3, 2019
Given the time differences between the US and Australia, my reply came a later (and I didn’t consult Randy or Martin before I sent it). Randy’s response had come overnight in Australian time.
When I had finished work for the next day, I wrote the following (Thursday, 3 October 2019 at 19:07):
Dear Greg (if I may)
Thanks very much for your enquiry and I think the ‘experimental’ approach to differentiating MMT from the mainstream monetary macro is interesting.
I laughed when I read your congratulations – we are definitely fellow travellers in the textbook sense – I never imagined writing such a book would be as difficult as it turned out to be. Much harder than a refereed journal article and/or a regular monograph.
Here are some ideas.
Randy, has provided one good example of the differentiation between the mainstream macroeconomics taught in our universities and the MMT approach.
Here are some more points of departure which can easily be framed as experiments, although in the real world we have already seen these examples play out before our eyes.
The first relates to the existence or otherwise of the money multiplier, which is a central mainstream proposition but is absent in MMT – in fact, we consider the causality runs in the reverse direction.
In mainstream courses, students get taught the money multiplier which says that the central bank determines the supply of loans in the economy by controlling the quantity of base money (reserves).
They learn that there is a (constant) ratio of broad money to base money, and when the central bank adds reserves, they are multiplied up to create a greater change in bank loans and deposits.
The theory has two elements:
1. Reserves constrain lending – a supply-determined approach.
2. Central banks control reserves.
MMT says that:
1. Reserves do not constrain lending – loans create deposits (not the other way around) – a demand-determined approach
2. The central bank does not control the amount of reserves.
3. Banks seek credit worthy borrowers – then create loans which create deposits.
4. The central bank will always supply reserves to banks seeking them – so unlike the Money Multiplier story – the base follows the broad (which is driven endogenously by demand for loans from credit-worthy borrowers).
5. The banks cannot clear an excess of reserves in the interbank market (this feeds into Randy’s point about the relationship between deficits and interest rates).
The causality is the reverse of the mainstream approach.
Thus, at the empirical level, mainstream macroeconomics struggles to explain why the broad monetary aggregates have not grown commensurately (as if multiplied) with the dramatic expansion of the asset side of central bank balance sheets.
MMT has no problem explaining that phenomenon.
Another example that follows relates to inflationary consequences of the expansion of the central bank balance sheets.
Mainstream theory predicted inflationary consequences arising from Quantitative Easing (QE) because they considered the increase in base money would feed in, via the money multiplier, to broad money, which, via the Quantity Theory, would drive up prices as bank lending accelerated.
MMT predicted that bank lending would not accelerate as a result of QE because the sluggish lending was not due to a deficiency of reserves (banks do not loan out reserves) but, rather, a shortfall of creditworthy borrowers due to the uncertain conditions following the GFC.
Further, MMT constructs QE as an asset swap only which may reduce interest rates in the relevant maturity range but has no direct implications for the growth of broad money.
Mainstream theory is problematic when trying to understand why the rapid expansion of central bank balance sheets did not cause inflation.
And relatedly, is the question of central banks and government bond yields.
Two points of differentiation.
1. Central bank purchase of government debt
Empirically, central banks have purchased significant proportions of government debt in recent years. The Bank of Japan now holds 43 per cent (approx) of all outstanding Japanese government bonds.
The ECB has a rising share as does the US Federal Reserve Bank.
Mainstream theory considers this to be ‘money printing’ and predicts it would be inflationary.
MMT indicates the inflation risk is in the spending choices of government (and firms, households etc) and not the monetary operation that might be associated with the spending.
In the standard mainstream government budget constraint rendering, governments have two choices to ‘fund’ their deficit spending: (a) bonds, (b) ‘money printing’. (a) drives up interest rates (and crowds out), while (b) is inflationary because it drives down interest rates and expands reserves.
MMT predicts that there is no difference in the inflation risk of the deficit spending arising from these ‘funding’ choices – noting that MMT does not consider the government to be financially constrained anyway.
So issuing debt does not reduce the inflation risk of the spending as the funds to purchase the debt were not being spent anyway.
2. Bond yields
Mainstream theory predicts that if the government increases its stock of debt as a result of running continuous deficits, bond markets will demand higher yields to compensation for the increased risk and higher expected inflation.
Take Japan as an example.
It has had continuously large fiscal deficits and QE programs which have resulted in a massive buildup of excess reserves in the banking system, yet interest rates and bond yields have been low (and sometimes negative) for several decades.
The empirical reality is the opposite to that predicted by mainstream macroeconomics.
MMT demonstrates that if the central bank leaves excess reserves in the cash system (as in the case of the Bank of Japan) then the interbank competition will keep the short-term rates low.
Further, MMT shows that the central bank can control all bond yields in any maturity range if it so chooses. This is exactly what the Bank of Japan has been doing for decades now.
Many financial market investors, following mainstream macro and monetary economics, moved their funds out of bonds and into real assets during the GFC because they expected yields to rise (and capital losses to be made) and inflation to accelerate.
They lost out on both accounts.
MMT predicted they would lose.
Think about the famous ‘widow maker trade’ on the 10-year JGB, which was based on mainstream theory predicting that yields would have to rise as the debt ratio rose.
The short-selling trades all lost out badly.
MMT predicted that the yields would not rise if the central bank chose to control them. The bond markets are supplicant in MMT rather than the drivers of yields.
And, how might mainstream economics account for the fact that:
1. Japan has the highest gross public debt to GDP ratio.
2. Yet, its interest-servicing ‘burden’ is about to be negative overall – that is, investors will be paying the government to hold their long-term bonds. This is because as positive-yielding assets mature, they are increasingly being replaced by negative-yielding issues, such that soon the overall debt servicing burden will be negative.
MMT explains this easily using the logic outlined above.
I hope these ideas help in your thinking about our work.
I could give more examples but time is short today.
I hope one day we might run into each other and have a chance to talk about these things in more detail.
All the best
October 4, 2019
On Friday, 4 October 2019 at 00:41, we received Greg Mankiw’s response:
Thank you for the quick replies. It will take me some days to think through all this, but I wanted to express my appreciate for your taking the time to address my question.
Which all seemed reasonable and I formed the view that he was considering our ‘experiments’ and would respond with his view about them, which would elicit further discussion.
That might seem optimistic to someone without direct professional experience in the Academy, but that is the normal way things pan out in my labour market.
Discussions occur between academics that never see the light of day and when they do are usually acknowledged in the credits of any paper or monograph produced.
Recognising that prior discussion and to-and-fro has occurred doesn’t imply that everyone agrees. But when an academic, as is standard practice, thanks others for their input, they are just recognising that there has been a discussion, or a person has read something or stimulated something that ended up in the final paper.
No more correspondence was entered into (as they say).
Then last week, he published his paper. I only was alerted to it on Twitter by an MMT follower.
Greg Mankiw had decided to have no further interaction with us for his own reasons.
In Part 2, we will discuss Greg Mankiw’s ‘guide’ or non-guide, as the case may be.
That is enough for today!
(c) Copyright 2019 William Mitchell. All Rights Reserved.