Misunderestimating MMT

My ex-colleague Gerard MacDonell wrote an article "The trouble with MMT" (Modern Monetary Theory). He offers a wide number of criticisms, however I believe that his substantive ones are unconvincing. I will note from the beginning that I cannot say that his arguments are wrong, they represent one side in a long debate in economics. Instead, I believe that we need to properly understand the debate, and that the MMT wing of post-Keynesian economics offers a lot of insight to that debate.

[Update: Gerard has a followup article here; I will respond to a couple of points in an upcoming short post.]

This article is partly aimed at people who are not familiar with MMT; Gerard's arguments are using concepts and language that are more familiar than MMT, and so I believe that understanding these arguments will be useful for many readers coming to grips with it. Another thing to keep in mind is that MMT did not just appear out of nowhere; it is a branch of broader post-Keynesian economics -- I discuss the relationship here. Since I view MMT as part of this larger tradition -- and not just the output of "MMT economists" -- it has much deeper intellectual pedigree than browsing some articles on the internet would suggest.

(This article dodges the discussion of political strategy that Gerard raises. My feelings on this are mixed. Although I believe in Functional Finance for an analytical strategy, the associated political strategies to justify programmes is more awkward.)

The Disappearing Fiscal Constraint

Gerard took aim at some of Warren Mosler's comments about the fiscal constraint. Since I view MMT as part of a broader tradition, we need to look at what that tradition suggests, and not quotes from a particular popular book. In this case, the MMT tradition relies heavily on Functional Finance, going back to Abba Lerner.

I will now restate my interpretation of what MMT says about the "financial constraint" for governments, using a fairly simple example.

Imagine that I walk into a new shopping mall, with only a $20 bill in my pockets. I can buy whatever I want -- so long as it costs less than $20 (sales taxes included). (I assume that I do not know any of the people running the shops, so they will not extend me credit.) I have a hard $20 constraint on my ability to finance purchases; hence it makes sense that I have a hard financial constraint. A key point to note that this constraint is directly related to a dollar amount, a fact that I will return to.

From a practical perspective, central governments face no such constraint -- the circular nature of governmental finance flows means that they can "finance" practically anything. (OK, under the current institutional framework, there are limitations; the U.S. Treasury could not just write a $100 quadrillion cheque to buy a fancy hotel on a Friday afternoon. I discuss such institutional details in Understanding Government Finance; What matters is that these limitations represent largely uninteresting "corner cases.") 

What stops the government from buying "everything"? Inflation. If the government issues an insane amount of liabilities ("money" and debt) relative to its tax base, the value of those liabilities goes down versus real goods and services. As a result, the constraint on government finances is best described: excessive government purchases and transfers can raise capacity utilisation in the economy beyond "normal" levels, (eventually) giving rise to inflationary pressures.

Since the government is buying stuff with money, is this not financial? If you want "financial" to mean "anything to do with money," I guess you could say that. However, this is a fairly meaningless definition. The constraint has nothing to do with the government's ability to finance purchases, nor does it have anything to do with fixed dollar amounts.

The second point (no relation to dollar amount) is critical, and why I think we need to drop the tag "financial."

I am currently developing my stock-flow consistent (SFC) modelling framework, and I have discovered that it is fairly easy to generate endlessly growing governmental debt levels. All you need to do is to add a sector to the economy, give it a source of cash flow, and forget to add an outflow. It will hoard financial assets, and drive the economy to a state of capacity underutilisation. The automatic stabilisers associated with fiscal policy kick in, and the government runs deficits to counteract the demand drag. Eventually, the economy returns to a "normal" capacity utilisation level, but with higher governmental debt and deficit levels. In other words, we can have wildly different governmental debt levels corresponding to the exact same level of capacity utilisation.

If we return to the real world, we do see sectors that emulate such hoarding behaviour.

• Pension funds are relentlessly accumulating financial assets to meet future outflows. Those pension funds will never purchase real goods and services (beyond their technology infrastructure), no matter the level of their financial asset holdings.
• Foreign central banks have been accumulating reserve assets, with no intention of ever buying real goods and services with those reserves -- they are to be used to purchase a particular financial asset (their domestic currency if it is under attack).
• Many corporations are mindlessly piling up cash hoards, reflecting the empire-building objectives of the C-suite.
• Increasing inequality is redistributing income flows to the parts of the household sector with a low propensity to consume, leading to the growth of the stock of financial assets held.
• If we include various intra-governmental debt holdings within the total, government debt outstanding could become arbitrarily large without any observable effect on economic behaviour.

As a result, we cannot relate the dollar amounts of governmental debt or deficits to real world capacity utilisation. Anything that causes such hoarding behaviour to increase will increase government debt level while keeping "capacity utilisation" essentially the same.

In summary,

• governments (that control their currency) can always "finance" spending (under plausible limits); and
• government spending is limited by an inflation constraint, and when that constraint is hit has no relationship to the dollar amounts involved.

Under such circumstances, my view is that anyone who applies the term "financial" to this constraint is using the word "financial" in a manner that bears no resemblance how that word is normally used by speakers of the English language.

With that detour out of the way, I will address what Gerard actually wrote. He appears to agree with the logic I outlined above, but he still thinks that the word "financial" applies. In the version that he has posted at the time of writing (Gerard extensively edits his articles, and so I am unsure what the reader may see in the future):

Finally, imagine that the debt/GDP ratio rises to a level that would otherwise be difficult to carry if funded with even one-week treasury bills.  The reason for defining the debt level in terms of what would be difficult to finance with bills will become obvious below. For concreteness only, call that debt level 200% of GDP.

Why I believe that he is wrong is that he is ascribing a direct relationship between the debt/GDP ratio and real variables ("200% of GDP"). His article makes no explanation of why this is possible, but I would argue that it is clearly wrong (for the reasons given above). For example, if we decide to capitalise actuarial liabilities and inflows onto the Federal government's balance sheet, the "debt" ratio could easily rise above 200% of GDP without it having an observable effect on anything.

The only way that I can make sense of Gerard's argument is to assume that the policy mix is unchanged forever, and so we can relate governmental liabilities to real goods and services in a more reliable fashion. For example, a 1% increase in the governmental deficit will imply a 1% increase in government consumption (which then feeds into capacity utilisation). I cannot guess why he believes this, but I would note that it follows from typical assumptions in mainstream economics. (To be clear, I am not saying that he believes all of the following points, rather this is my characterisation of standard mainstream views, which appear to be consistent with his analysis.)

• It is typically assumed that there is only one representative household (hence no income distribution effects);
• there is only a single composite good (which allows for a well-defined production function with a single capacity variable);
•  the business sector is a cipher that does not accumulate financial assets; and
•  models typically do not include foreign central banks that hoard reserve assets. 

In a model that embeds these assumptions, we would see a simple relationship between governmental deficits and capacity utilisation. However, since I view those assumptions as dubious, I do not expect such a relationship to exist.

[Update: added this paragraph.] This discussion relates to the MMT/Functional Finance view of financing as it relates to analysis; there is no claim that eliminating the "financing constraint" from analysis opens space for new policy. For example, the government still needs to impose taxes, even if they are technically not "financing" spending. (The MMT view is that taxes drive the value of the currency -- Chartalism.) However, it does eliminate analytical errors that drive policy mistakes. For example, the Maastrict Treaty 60% debt-to-GDP ratio made no sense, nor did the search for the magical 90% debt-to-GDP ratio that appeared in the more lurid corners of economic analysis. In other words, MMT analysis does not offer the prospect of a free lunch (disclaimer follows) -- but it helps avoid wasting food as a result of ignorance. (The Jobs Guarantee can be viewed as providing a free lunch -- the programme offers the possibility of eliminating underemployment while at the same time offering a more humane method to anchor the price level.)

MMT and Money Financing

Gerard then turns to the question of money financing. He discusses two possibilities:

1. Excess reserves pay the policy rate of interest; and
2. excess reserves do not pay interest.

He wastes an incredible amount of space on the first possibility -- he apparently missed the part of the MMT proposals that emphasises that the second possibility is what is in mind.

Unfortunately, by emphasising the first, he did not take the second seriously. Highlights of his comments:

I don’t want to get into this scenario too deeply because it is so unrealistic and because it is quite a bit more complicated than I am smart.  But we can imagine two ways in which this case quickly goes intolerably inflationary, one following a conventional logic and one following the fiscal theory of the price level. I will be brief – and yes, conclusory – on each.
...
If an interest rate of zero is too low relative to how the economy is doing, then we get inflation as the Fed falls increasingly “behind the curve”, by virtue of not being able to move.  I will stop there because the story is very familiar to all of you, and I have nothing to add to it.

Essentially Gerard is assuming that locking the policy rate at 0% will cause inflation, because the central bank can no longer do anything with interest rate policy. The basis of this assumption: "it is very familiar to all of you."

Ahem. The effect of interest rates on inflation is along running debate, and was not invented by mainstream "neo-Ricardians." (It only seems that way because they ruthlessly exclude citing anyone other than other mainstream economists.) The MMT position on locking the interest rates at 0% hinges on the argument that interest rate policy has only a limited effect on inflation. If you want to disagree with MMT, you have to take on that argument.

I am agnostic on the question of the effectiveness of interest rates (it's an open debate within post-Keynesian economics, and even within MMT). That said, the modern mainstream literature that I have seen is entirely useless. The effectiveness of interest rate policy is assumed within the mathematical structure; it is literally impossible to debate the topic using DSGE models. The theory is non-falsifiable.

How would we falsify the theory? We would need an episode where the central bank holds the policy rate away from the natural rate. The recent period of zero rates in the United States would have qualified -- if we used pre-crisis estimates of the natural rate. However, the mainstream estimates the natural rate of interest by using techniques that assume that it has the assumed effect on the economy, and so the natural rate was conveniently estimated to be negative, so that the theory appears to predict the lack of inflation. (I discussed this topic at greater length here.)

I am not saying that Gerard is wrong, rather that I see no evidence that his view is right. As a result, he is not in a position that the MMT view is wrong. In any event, I do not see the proposal to lock interest rates at 0% as being a core part of MMT theory.

(He later discusses the Fiscal Theory of the Price Level, which I have discussed elsewhere.)

Concluding Remarks

Gerard MacDonell's article illustrate some of the difficulties MMT faces for broader acceptance. In addition to difficulties with terminology, there is a constant need to discuss how underlying assumptions differ from mainstream economics. Having to fill every single article with discussions of theoretical concepts so as to avoid being misunderstood when being quoted out of context makes writing awkward.

(c) Brian Romanchuk 2017