Recent Posts

Wednesday, March 4, 2020

Initial Comments On Fullwiler's Fiscal Sustainability Paper

Scott Fullwiler's article "The debt ratio and sustainable macroeconomic policy"* offers a very good introduction to Modern Monetary Theory's (MMT) stance on fiscal policy. It covers a lot of ground, making it hard to summarise. This article has some general remarks on why I recommend reading the paper as an introduction to MMT thinking. The key argument is that interest service is what matters for sustainability (and not the debt-to-GDP ratio) -- and that the rate of interest is under the control of governments.

Introduction

My current (new) writing project is to write a chapter on fiscal policy for my MMT handbook. The projected contents appear to parallel the Fullwiler paper. From my writing perspective, the difficulty is that I want to break the chapter into sections that are tightly focussed on particular topics. The result is that discussion of the Fullwiler paper would make its way into each section. This messed up my plan to do a deep dive on the paper as a single section.

One initial point of interest is that if one is approaching the article from a background as a practitioner in finance, the analysis does not appear very radical. A practitioner might not agree with it, but the discussion has nothing to do with the strange stereotypes about "printing money" that popular critiques of MMT focus on. If people actually read the journal articles, MMT discussions would end up being rather boring.

Contrast With Neoclassical Fiscal Theory

One area where I have some disagreements with the Fullwiler text might be in the treatment of neoclassical ("mainstream") theory. Since he is contrasting and comparing the MMT view with mainstream views on a number of topics, this is somewhat inevitable. My concern is that I doubt that we can usefully characterise neoclassical thinking on fiscal policy.

The term "neoclassical economics" appears vague, but we can give a fairly rigorous definition for which journal articles meet that criteria.
  1. The use of optimising mathematical models often appears.
  2. When we look at cited papers, neoclassical papers almost exclusively cite other neoclassical papers. One could use graph theory to characterise the linkages (but it has been decades since I last looked at graph theory, so I will not pretend to do so). By contrast, heterodox authors will cite both other heterodox authors and neoclassical papers.
However, this ends up only being a definition of a group identity. If we look at the theory around fiscal policy, there is no real consensus on what the theory says. This is unlike other areas, where there is more of a consensus. The end result is that one can point out dubious statements made by particular neoclassical economists, but they may be speaking only for themselves -- and they may have no settled theory to justify their claims (if they are writing an opinion piece, for example). This means that "contrast and compare" exercises are a bit of a snark hunt, and why I think it is best to focus more on the MMT contributions.

Sustainability

I will now turn to the Fullwiler article. It starts off with the following statement:
There are few issues of more theoretical, empirical, and political interest than fiscal sustainability. This is not surprising, since in addition to its own significance it is at the core of so many other fundamental debates, such as the ability to pay future entitlements, central bank independence, “printing money,” and the appropriate role of government itself in the economy.
To what extent a consensus exists in neoclassical thinking on fiscal policy, invoking the term "sustainability" seems to be the commonality. This explains the importance of the topic. The problem is that "sustainable" is not really a well-defined term from the neoclassical perspective.

Fullwiler describes sustainability as follows:
So, importantly the sustainability of the government’s fiscal position is not about the government’s ability to spend by crediting bank accounts – though this is very important for understanding why a government can always “afford” policy actions that enable a full employment economy and why it can never be forced into involuntary default via an inability to pay or service its debts – as much as it is about the size of the debt service relative to the size of the economy.
It appears that this is the definition of fiscal sustainability from the MMT perspective (although one could probably find statements that are effectively equivalent).

One could reasonably argue that this should be equivalent to theoretical definitions of "sustainability" used by neoclassical theory. My argument (and it is possible that nobody agrees with me) is that this does not fit the facts on the ground. When we look at opinion pieces written by neoclassical authors, we can see a few variants of "sustainability" that are effectively being used.
  • Sometimes, the above definition is used. In which case, there is no fundamental disagreement.
  • An invocation of the "inter-temporal governmental budget constraint."
  • Any rise in the debt-to-GDP ratio will be described as "unsustainable," on the argument that it cannot go to infinity. Therefore, fiscal policy has to be projected to keep the debt-to-GDP ratio at some arbitrary level on some arbitrary horizon. No explicit notion of debt service appears, it is at best implicit.
  • There are tipping points in the debt-to-GDP ratio, where the government is vulnerable to runs by bond vigilantes, or reduced growth. Less popular in 2020, but still pops up on occasion.
  • Obviously incorrect models -- e.g., CBO projections (see below) -- are used to make statements about trends in debt service and debt ratios.
  • In popular discourse, back of the envelope estimates about the taxes needed to pay for programmes, and then the Laffer curve is invoked. (This was recently done in a very shaky budget analysis.)
Outside of the governmental budget constraint, the key is that the definitions are somewhat vague. This is what allows for the fluidity of neoclassical analysis of fiscal policy -- it is always possible to find neoclassicals on either side of a fiscal policy proposal.

The Governmental Budget Constraint

The following discussion represent my views, and not based on the Fullwiler article. However, it needs to be covered, given that it often pops up in more theoretical discussions.

The intertemporal budget constraint is an equation (inequality) that appears in some neoclassical models. (Some people will also include an accounting identity as being part of the budget constraint, which is confusing.) The constraint is that the net present value of government debt has to go to zero as time goes to infinity (which is mathematically equivalent to another statement about primary budget purposes).

This does not mean that government debt has to go to zero as time goes to infinity, rather its growth rate has to be less than the discount rate.

If the constraint appears in the model (not all models feature the constraint), it limits the scope for fiscal policy in the long term, although it tells us nothing about behaviour on any finite horizon (a point that is typically ignored).

From an applied mathematical perspective, this constraint is not dealt with a satisfactory way.
  • There is no good** reason for the constraint to hold.
  • Since it is an assumption, it is literally impossible to model what would happen if a government tried to break the constraint within the model.
  • I have never seen an attempt to see the effects of what would happen if fiscal policy tended towards the limit of the constraint (outside of the heretical Fiscal Theory of the Price Level).
The mathematical concerns translate into real world problems. The theory offers explanation of what happens if a government just decided it wants to ignore the constraint, nor any good idea how a government is supposed to ensure that policies conform to it.

Although the constraint has attracted the ire of heterodox authors (including myself), its lack of practical significance means that it does not come up in most discussions of actual policies. To a certain extent, it only seems to show up as a bit of mathematical snobbery ("So what is the governmental budget constraint in your model?")

r, g and all That

The bulk of the article focuses on the conditions required to keep interest service contained. I will return to that subject in more detail later, but  just want to touch on a key point: the critical importance of the relationship between the interest rate on government debt, and the long-term growth rate. (These are normally defined in real terms due to the theoretical preoccupations of neoclassicals, and so they are referred to as r and g.)

The basic premise is that we drop the bizarre insistence that the net present value of government debt tend to zero to an economically literate belief that the debt-to-GDP ratio remain finite. What we see is that if the interest rate on debt is less than the growth rate of GDP, it is extremely hard for anything exciting to happen to the debt-to-GDP ratio: economic growth compounds faster than the debt.

We then arrive to a key MMT divergence from the mainstream: that interest rates are a policy variable largely determined by governmental policy. In particular, there is no tendency to drift towards a "natural" rate of interest. If we accept this, it is clear that it is pretty easy to keep interest rates below the growth rate of the economy (other than in the rather implausible scenario of a hyperdeflation).

The article does an analysis of the current situation, as well historical data. What we see is that the high real interest rates that happened in the aftermath of Volcker's crushing of the economy was an outlier. However, neoclassical thinking extrapolated that experience as being a natural state of affairs.

(The prospect of negative interest rates is an added body blow to worries about the debt-to-GDP ratio. If interest rates are negative, ramping up debt levels increases government revenue.)

We see that the key theoretical dispute between MMT and conventional economics here is about the properties of interest rates, and not fiscal policy itself. The difference is large -- but is almost never addressed in most critiques of MMT.

Botched Models

One of the problems with sustainability discussions is the willingness of mainstream economists to work with models that obviously do not work. Fullwiler:
In other words, a primary budget balance not at least as high as 0.6 percent of GDP on average would grow deficits, the national debt, and debt service all to the point that eventually paying the debt service would result in high and rising inflation. While this second row puts the convergence ratios at infinity for convenience, in fact at some point the increased debt service would simply pass through to inflation to raise nominal GDP in kind. Thus, CBO’s [CBO = Congressional Budget Office] regular practice of assuming a long run nominal GDP growth rates equal to the potential real GDP growth rate plus inflation at around 2 percent is inconsistent with its own projections of unbounded growth in debt service payments

In other words, they assume a growth rate that is independent of what is actually happening in the economy. The debt-to-GDP ratio only goes to infinity because the model is internally inconsistent.

If we look at stock-flow consistent models, we can see that the entire premise of the debt-to-GDP ratio going to infinity makes no sense. Something else would give first, since entities are unlikely to sit around with arbitrarily large amounts of financial assets. As a result, it is puzzling why anyone would raise the possibility, yet one sees it in internet debates.

Ricardian Equivalence, and So Forth

The article also discusses the notion of Ricardian Equivalence, as well as the widespread belief that bond issuance is less inflationary than "money printing." For reasons of space, I will discuss this in a later article. However, this is yet another area where there appears to be less of a consensus in the mainstream.

The arguments about bond issuance being used to reduce inflation is quite well known, and pops up continuously in internet articles and op-ed pieces. It appeared in some Economics 101 textbook, and has taken a life of its own. (The only introductory mainstream textbooks I own do not contain it, so I do not know where it comes from. My collection of mainstream texts are at the graduate level, and they do not contain this concept.)

The problem is that the idea that bond issuance is anti-inflationary makes no sense, either from the MMT or modern neoclassical perspective. Money demand is a preference of the private sector, and within the framework of an inflation-targeting central bank with an interest rate as a policy lever, money holdings are outside the control of the central bank. (Even when central banks allegedly targeted the money supply, the practice was to set interest rates to keep money growth near target -- which is similar to how an inflation target is implemented.)

Quantitative Easing appears to be an exception, but this is not the case. The reason is that the central bank needs to pay interest on excess reserves (settlement balances) at the overnight risk-free rate (modulo tiny spreads). From a practical and theoretical perspective, those excess reserves are indistinguishable from overnight Treasury bills. The only way the difference in instruments matters is in the discussion of a potential default, since those are central bank liabilities rather than that of the fiscal arm of government.

The issue here is that beating up on this belief is somewhat of a distraction, since it is only held by the (rather large) community that thinks that Economics 101 textbooks are the culmination of economic thought.

Concluding Remarks

The Fullwiler article offers an extremely readable introduction to MMT thinking on fiscal policy. I will return to some of the concepts in later articles. The key take away is that the notion of sustainability is tied to debt service, and the main point of differentiation between MMT and neoclassical thinking is the argument that long-term average of interest rates on government debt is under control of policymakers (and not set by some law of nature).

Footnotes:

* Fullwiler, Scott T. "The debt ratio and sustainable macroeconomic policy." World Economic Review 7 (2016): 12-42.

** The story is that the constraint comes from a transversality condition on the household sector budget constraint. This ignores the reality that there are no transactions by which the household sector can affect the aggregate amount of debt outstanding. Nevertheless, I have never seen an attempt to justify the condition in any of the many articles I read, it is just asserted to hold.

(c) Brian Romanchuk 2020

5 comments:

  1. "The problem is that the idea that bond issuance is anti-inflationary makes no sense, either from the MMT or modern neoclassical perspective. Money demand is a preference of the private sector, and within the framework of an inflation-targeting central bank with an interest rate as a policy lever, money holdings are outside the control of the central bank."

    The divergence of theories is so wide that conversation is nearly impossible. Take my mechanical theory for example:

    Mechanically, "money demand by the private sector" is a vague and unmeasurable concept. Everyone would like to have more money than they presently have but the money supply is limited. Paradoxically, mechanically speaking, money supply is not limited because government can increase the money supply at-will.

    As a mechanical result of this paradox, theorist need to look to the concept of "control". Who controls the supply of money? Well, mechanically, the private sector controls money once it has been issued. However, government controls money to the extent that government can tax AND to the extent that it can issue NEW money. "Control" is the dominate concept.

    Government has control, relinquishing control (to the private sector) to some politically determined extent.

    Now I understand that the mechanical approach just outlined is not part of any classical framework that I am aware of. That places me in the position of arguing with everyone coming before me. Which makes it nearly impossible to have a theoretical conversation.

    ReplyDelete
  2. Hi Brian,
    Take a look at a paper Jamie Glabraith wrote about this in 2011. It's at http://www.levyinstitute.org/pubs/pn_11_02.pdf
    After reading his paper I had some fun reproducing the graphs showing unsustainability of debt service at various times in the future by varying the interest rates as the CBO does. It's complete nonsense of course. The models are internally inconsistent as you say, and demonstrates that the real issue is about justifying austerity today.

    ReplyDelete
  3. How about a link to Fulwiller's paper? Or perhaps there is one above: I'm very good at staring straight at things and not seeing them...:-)

    ReplyDelete
    Replies
    1. I probably should have added it, but wasn’t sure how stable the link would be. I found it via Google Scholar. For now, I’ll let people search for it that way.

      Delete

Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts.

Note: if you want to post comments from Apple devices (iPhone, iPad), you apparently need to turn off "prevent cross-site tracking" in Safari privacy settings. (The reason presumably is that another URL handles comments, and so the user session needs to be preserved when redirected to that site. I don't like this, but this is not enough to make me switch my hosting service.)