Pages

Wednesday, March 13, 2019

What Did We Learn From The MMT Maelstrom?

My guess is that every prominent economist that wants to mouth off about Modern Monetary Theory (MMT) has already said their piece, and the arguments that swamped my Twitter feed may finally go back to their earlier levels. It is abundantly clear that the prominent New Keynesians who attacked MMT want it to disappear, but that seems unlikely (although I am obviously biased in that assessment). From the perspective of the history of economic theory, the neoclassical reaction was following past form, and we can expect the same patterns in the future.

The Maelstrom

The early months of 2019 saw a flood of attacks on MMT, with almost all the articles following into the following types.
  1. Prominent neoclassicals presenting straw man versions of MMT and attacking those versions. Arguments were presented that MMT is "just" IS/LM. (This is not a new tactic, as I discussed this before.) Despite having top tier academic credentials, these luminaries are unable to provide any useful references from the MMT literature. 
  2. Left-wingers who otherwise have a negligible following have discovered that they can gain temporary prominence by attacking MMT. 
  3. Right wingers writing articles screaming that "MMT is socialism" and that hyperinflation is around the corner. (Admittedly, many of the prominent New Keynesians argued the same thing about hyperinflation.)
Since none of these attacks on MMT actually discussed MMT (with the exception of reviving some old internet arguments from the dead), there is literally nothing worthwhile to respond to. Instead, I will discuss what we can learn from this episode.

(On the chance that the reader wants to get a sample of these criticisms as well as a response, this article by L. Randall Wray offers some of the highlights.)

Update: I was Early (by a Few Hours)

My call for "peak MMT madness" was off by a few hours; we reached the peak this afternoon (Montreal time). I do not apologise for being early: researchers need to inform clients of the trend change before it happens, so that they can position themselves accordingly.

The peak was the publication of this survey on MMT by the IGM Economic "Experts" Panel. In my view, it will be nearly impossible to upstage this piece of work.

The IGM describes these experts as follows:
For the past two years, our expert panelists have been informing the public about the extent to which economists agree or disagree on important public policy issues. This week, we are delighted to announce that we are expanding the IGM Economic Experts Panel to add ten new distinguished economists. Like our other experts, these new panelists have impeccable qualifications to speak on public policy matters, and their names will be familiar to other economists and the media. 
As for the survey on MMT, I leave the contents for historians of economic thought to judge. Any attempt to describe the current situation in economics will need to match the assertion that these are distinguished economists versus the questionnaire contents. To be fair to the economists surveyed, they were not responsible for the questions, and some noted that questions were ambiguous.

I draw the reader's attention to the second question:
Question B: Countries that borrow in their own currency can finance as much real [emphasis mine] government spending as they want by creating money.
If the premise of that question were correct, it would be possible for the United States Government to purchase eleventy quadrillion aircraft carriers in 2019. Needless to say, that is physically impossible, and most reasonable people would point out that the person asking the question has gone completely nuts. To be completely fair, at least one respondent noted that point, but it should have been completely obvious to everyone involved. (One might argue that the survey questions were probably written up by some intern or grad student, and they might not have expertise in macro, and so did not understand the implications of the question. However, it seems completely unlikely that there was nobody senior at the IGM who is taking credit for the surveys, and somebody who knows at least something about macro had to have fed some information to the unfortunate survey builder.)

No competent historian of economic thought would actually impute that view to MMT. If the word real in the question was changed to nominal, you would run into the MMT position: if the government wants, it can outbid everyone for the finite stock of real goods and services on offer -- with the side effect of driving up the prices for those goods.

If anyone wants my opinion on the intellectual integrity of economics, I think they need to think deeply about the implications of this survey.

Keep Doing What I am Doing

I largely confine my writing to a narrow segment of macroeconomics: the parts that rates market participants care about. This spreads out into a lot of topics of interest to others -- i.e., the limits of fiscal policy -- but such readers are more interested in the analysis of policies that are being proposed, not proposing new ones themselves.

For such a reader, this whole episode is just an embarrassment for the so-called authorities among the neoclassicals. If one is interested in the theory, one does not care about straw man attacks on it. The only question that really matters: is it a theory that is progressing, or regressing? I am writing about MMT for a reason, and that is the reason. I am currently researching a book on recessions, and unlike the various New Keynesian luminaries, I am going to give a good faith summary of the theories I am not a fan of. Unless I run into some gems of new research -- which I accept as a possibility -- neoclassical theory has severe drawbacks when compared to post-Keynesian approaches in that area. 

I saw three common lines of criticism of MMT, which I may keep in mind when writing, and the reader might be interested in pursuing. As I will explain in detail, I think these arguments are misunderstandings of the actual situation.
  1. Does MMT have empirical work?
  2. Do MMTists question their assumptions?
  3. Does MMT have formal theory in the form of mathematical models?
I will discuss these in turn.

Empirical Work


One of the reliable stylised facts about economic discussions around MMT is that a condescending neoclassical graduate student will appear to announce that MMT has no empirical work. This is the result of The Iron Law of Research: if you do not read the literature, you will not find the research.

When I was first introduced to MMT, even though I was a mere control systems engineer looking at economic theory in my spare time (day job was rates analyst), I managed to find Full Employment Abandoned by William Mitchell and Joan Muysken. It is a mixture of theoretical and empirical analysis of trends in the labour market, and offers a compelling alternative explanation to the collapse of inflation in the 1990s. If I could do that, what exactly are they teaching neoclassical doctoral students in economics nowadays?

There's certainly more empirical work; there's at least one in the L. Ranndall Wray article I linked earlier; one can search the resources at neweconomicperspectives.org. Since MMT covers a lot of ground developed over decades, there's a lot of empirical work already done. It's up to the reader to find work of interest.

The MMT-oriented institutions do not have the funding that is showered on neoclassical economic research; for obvious reasons, the pure numerical statistical work is going to be done at "mainstream" institutions. The post-Keynesians have access to the very same statistical packages as the mainstream, and so many of the tools are the same. The methodologies differ, for reasons that are well articulated by the heterodox community.

I am applied mathematician; I do my own empirical work, thank you very much. Since nobody wants to read boring stuff about statistical tests, I shy away from that. For the very same reason, MMT primers on the internet don't attempt to lose readers by regurgitating statistical mumbo-jumbo. The "empirical" content will usually cover the cases where the neoclassicals were hilariously wrong -- which is extremely empirical, just the neoclassicals want to ignore those episodes.

I have eccentric views on economic models. They are:
  1. You cannot do good empirical work in macroeconomics without reference to a good model.
  2. All macro models are terrible. (My stepped up version of the "all models are wrong" mantra that various mainstream economists intone, without actually thinking about the implications.)
Put those two points together, and guess what? One ends up with profound skepticism about academic empirical economics. I do not think I was completely alone in that view, although I was probably an outlier in terms of awareness of the issue. I both produced and consumed interest rate research for about fifteen years. On paper, the mainstream empirical research ought to have been up the alley of rates market participants. In reality, very little was looked at. My explanation is straightforward: academic/central bank research is judged on its ability to impress peer reviewers; market research is only judged on its ability to make money. Making money is far more objective and stern criterion than vomiting the results of statistical tests that are chosen by convention.

To repeat: these are my views, and not "MMT's". That said, I am far more sympathetic to the post-Keynesian approaches to empirical work than the mainstream's as a result of those prior views.

Finally, there are questions about empirical testing of MMT proposals. If one actually read the MMT literature, the argument is made that the Old Keynesians as well as the New Keynesians approached fiscal policy incorrectly with aggregate demand management. The models relied on aggregated behaviour, and the hope was that a rising tide would lift all boats equally. This is exactly what did not happen, and so inflationary bottlenecks crop up. To fairly do an empirical analysis of MMT policies, we would need a data set of countries where policymakers listened to MMTists, not neoclassicals.

Questioning Assumptions


Although not repeated as often, I saw a few variants of the criticism that MMTists do not question their assumptions, or less politely, "a cult." The usual phrasing invokes Scientism: it's scientific to question your assumptions!

Having gone through an undergraduate degree in the applied sciences, I would respond that these representations of the sciences are highly questionable. Undergraduates are taught year after year with the same texts. We do not see Circuits I professors stopping in the middle of lectures and asking: what if Kirchoff's laws are wrong? Even in my field of control systems, I would have been happy teaching an intro course to signals and systems using a textbook that was decades old -- even though the entire research and advanced design paradigm had shifted. The internet/op-ed arguments about MMT are on basic issues, for which no new information has arrived for decades. For more esoteric topics, research is ongoing.

However, to understand MMTist behaviour, we need to step back and think about the context. Modern Monetary Theory fits within "broad tent post-Keynesianism," as described by Marc Lavoie in Post Keynesian Economics: New Foundations. (There are a few "narrow tent Post-Keynesians," who dislike MMT, but tough luck for them; nobody is going to come up with another label for "broad tent post-Keynesianism" to please a shrinking group of zealots.)

That is, MMT is a sub-set of post-Keynesian economics. There's a few constraints on a theorist to stay within that set; probably the most important of which is an affinity to a floating exchange rate. If a theorist questions a few assumptions of MMT, they just end up being classified as another branch of post-Keynesianism. Why should anyone care about that?

It takes a lot of effort to move from inside the post-Keynesian set to an internally-consistent set of beliefs. I've only really encountered the following alternative (internally consistent) views:
  • neoclassical economics, which is split into different genres;
  • doctrinaire Marxists (I think some Marxists sneak into the edge of post-Keynesianism);
  • Austrians;
  • physical scientists pushing mechanistic models of the economy.
If you look at this list, you realise that in order to not be a post-Keynesian, you have to reject almost everything in post-Keynesianism, which is far beyond questioning an assumption or two.

If one reads between the lines of Lavoie's book -- which lovingly details the dozens of factions within broad-tent post-Keynesianism -- one can see that post-Keynesians managed to rip themselves to shreds in completely pointless theoretical disputes. (My Twitter timeline used to be filled with such attacks before they became so deranged that I quietly did a mass unfollow.) Given that behavioural pattern, one is perhaps not surprised by how post-Keynesians were marginalised in the overall debate on economics. The "MMT project" dropped the emphasis on the petty infighting, and instead focusing on advancing a theoretical/policy agenda. As such, MMT is self-selecting: whiners who want cry about minor details select themselves outside of the group. More politely, MMT is characterise by "joiners," not "splitters" (I think that Marc Lavoie used that terminology in a different context). Such cohesive behaviour is not unprecedented: by forcing macro research to use the DSGE paradigm, the mainstream selected out the heterodox elements.

Formal Theory


In my lengthy experience, people fetishize mathematical models in economics and finance. In the case of finance, the mathematical models have an important place, but the reality of mark-to-market accounting tends to constrain one's enthusiasm. In economics, the inherent difficulty of forecasting really hammers the simplistic nature of mathematical models.

There are plenty of mathematical models in the MMT literature; they are almost all (I believe) stock-flow consistent (SFC) models (and perhaps some toy models that might be special cases of SFC models). These models are mathematically simple, but the simplicity is based on some very cagey analysis of the limitations of mathematical models in macroeconomics. Speaking as an applied mathematician who learned economics on the job, the wrong place to understand SFC models is in complex models in the literature. Instead, you need to start with the basics, and learn why the models are constructed the way they are.

I would argue that Godley and Lavoie's Monetary Economics is the go-to textbook for that. My book on SFC models in Python offers an inexpensive, breezy introduction to the concepts, but I would argue that it is best a sampler for the delights found in Monetary Economics (unless you want to use my Python sfc_models package, as it is the de facto user guide).

The key is that you do not fall into the trap I have seen others fall prey to: you cannot strip out the mathematical model and then ignore the text around it. These models have limitations -- remember, all models are terrible -- and the eye is immediately drawn to those limitations. In most cases, the key is not learning how SFC models work, rather it is unlearning your prior assumptions about mathematical economic models.

Obviously, people will want to debate the models. However, that requires that they actually roll up their sleeves and read the literature.

Finally, I want to underline that we cannot capture all of MMT with mathematical models. For example, one of the more interesting parts of MMT research for rates market participants is the legal analysis of default, monetary operations, and the framework for central bank independence. If we want to use mathematical models, we cannot fit them to what we want to happen, they have to match the legal reality we live in. In reality, floating currency sovereign default is hard to arrange within most (all?) institutional frameworks, and we need to listen to legal opinions, not those of finance bros.

Room for a Neoclassical-MMT Synthesis?

Anyone coming to the horror show of academic economics from another field has to wonder whether there is a better way. Anyone with an applied science background would be very tempted to rip out all the mathematical models and empirical work out of the existing body of knowledge, and start over.

I have sympathies for that view. The creation of a collective group that writes a self-contained economics text that publishes under a pseudonym and similarly ignores claims of academic precedence might solve a lot of problems. However, that is easier said than done: the theory determines what are acceptable models and empirical methods.

In particular, one might ask: can neoclassical economics swallow up MMT? Certainly, one could use neoclassical arguments to justify the same policy proposals as MMT, but that just tells us about the theoretical ambiguity of neoclassical macro. Since the leading lights of neoclassical economics refuse to even cite MMT papers, one could easily imagine them claiming to have covered "MMT" with some neoclassical thinking.

However, for anyone who is intellectually honest, swallowing MMT would be pretty awkward. If we look at the actual contents of MMT papers, they are filled with discussions how the underlying assumptions of neoclassical models are outright incorrect. That's going to be pretty hard to reconcile.

Even more limited contributions will not be embraced by the entire neoclassical community. The fear of "bond vigilantes" and "the 90% debt limit" were already laughable within the context of neoclassical models; they did not need MMT to reject such thinking. However, making up vague stories about debt limits is too politically useful for neoclassical economists as a class to give up. We end up with the strange situation of neoclassical economists disclaiming any responsibility for the policies followed by other neoclassicals (e.g., austerity). Was this not supposed to be a rigorous, mathematical, science? This shiftiness turns neoclassical economics into a giant game of Calvinball.

Furthermore, it is very hard to see the central bank-oriented neoclassical consensus embracing the actual contents of MMT thinking about monetary policy. I admit that I am perhaps too cynical, but I cannot see how embracing a view that interest rate policy is possibly ineffectual -- and possibly with the opposite sign assumed by convention -- is anything other than a career-limiting move in a central bank. Perhaps I could be proved wrong, but I will believe it when I see it.

Concluding Remarks

For those of us who are interested in economic theory as theory -- and not a battleground for political economy -- the recent uproar told us nothing.  (I am interested in political economy, but as usual, I am an eccentric on that front as well.) There are good faith arguments about MMT, but you need to read the damn literature in order to discuss those.

I hope that this is the end of my comments on the internet/media brouhaha. I have been waiting (for a long time) for the "big" MMT macroeconomics textbook before I turn to write some primers on topics of interest. These primers would hopefully answer some of the legitimate theoretical questions people have (including myself, as I have a hard time differentiating "MMT" from "post-Keynesianism," as I do not make a big deal about distinguishing the two).

NOTE: The book covers that appear in my articles are affiliate links to the books on Amazon.com; I get a small fee if you follow them and make a purchase. The three books by other authors are highly recommended if you want to get a handle on these topics. Once again, I am book-oriented in my learning, not article-oriented. 

(c) Brian Romanchuk 2019

15 comments:

  1. If the MMT people were operating under an empirical method you would see them constantly making adjustments which I dont see them ever doing...

    I cant think of an adjustment they have ever made..

    They are just advocating for their own Theory which goes the same for the others who are opposing them.. best result to expect out of all of this for the MMT side is some sort of synthesis between the opposing theories...

    ReplyDelete
  2. Think of a control system where the feedback circuit compares as non zero error value from the set point and no adjustment to the control variable is ever made... doesnt work.

    ReplyDelete
    Replies
    1. I taught control systems, and like I wrote, I would have no problem teaching an introductory level version from a textbook that was written before the paradigm shift that hit the field in the early 1980s. (From a practical perspective, the only issue would be modernising the assigned problems to match the practice of relying on digital computers.)

      The level of most MMT discussions on the internet are introductory level; there's not been a huge wave of information (since the late 1990s) that invalidates those basic views.

      Delete
  3. Well, I'm not concerned with my terrible interpretations of the wonderful Godley and Lavoie models making it into an academic journal any time soon, I am, however, keen to roll up my sleeves in an attempt to learn something about how real world monetary systems actually work.

    I know you’re not a fan of agent-based models, nevertheless, this is the route I have chosen. Next up will be ModelPC!

    https://www.tenjinmacro.com

    Thank-you Brian for another interesting blog post.

    Regards,
    Dan.

    ReplyDelete
    Replies
    1. I'm equal opportunity; I'm not a huge fan of any class of economic models...

      Agent-based models seem interesting, but my view is that it should be possible to get them to fit any desired behaviour. That is, so long as accounting constraints are respected, we can back fit behavioural rules to generate any pre-determined outcome. (This may not be true for models where there is strong herding behaviour, but that raises the issue of why would expect competing entities to herd. Most industries are moving towards oligopoly status, and so aggregate behaviour in that sector is the result of a few central planners. If the agent-based model respects that reality, we end up that behaviour is the result of a few competing central planners, which creates a pretty wide range of potential outcomes.)

      Agent-based models might work very well as teaching models, but the flexibility of outcomes should make it hard to draw conclusions.

      Delete
    2. Agreed, the flexibility of outcomes might make it hard to draw conclusions, but, in the simplest monetary model seeded with real-world (expenditure) stimulus time-series (for gov sector) and accounting restraints respected, other sectoral agent outcomes are pinned down.

      Looking at the model net flow of financial assets to consumer agents (real world UK and US time-series), there is a similar story of decreasing flow in the years preceding every real world recession.

      Of course, one would have to take into account real world private sector credit actions. But, none the less, interesting...

      Delete
    3. Right now, I’m a bug on recessions. The interesting class of recessions is the ones where there is a collapse of fixed investment that coincides with a shutdown of the credit markets. I can see how it would be easy to generate such an outcome in an agent-based model, which is useful. However, my feeling is that it could be attacked on the grounds that the model just builds in a prior assumption; the “reaction function” that I impose on credit investors bakes the outcome into the cake. Shouldn’t credit investors be looking for “optimal equilibria”?

      (I’d be happy to get my hands on appropriate research from the literature if it’s up that alley; haven’t gotten to that point yet. I’m not looking forward to building my own version of such a model, which makes working it in harder. I’ve got my Python SFC models framework ready for work like that.)

      Delete
  4. Brian Romanchuk, I understand your frustration with regards to mainstream economists misrepresenting MMT, but these people do make serious efforts to say stuff that's not dumb. Could it be that MMT is continuously being misrepresented by leftwing politicians as a license to spend (by the likes of AOC)?

    Even among the scholarly MMT enthusiasts there has been an internal marriage with the Green New Deal.

    Frankly, I think as soon as you recognize that the balance sheet of the central bank and the debt of the government is part of the same (since the CB will in the end support the gov no matter what, and gov don't seem to care about deficits anymore), you see the net gov debt of the US (and europe and especially Japan) is actually much lower, and much more importantly, Central Banks already know that it is inflation that sets the limits to spending, and little else.

    The Central Bank & US gov tandem is already doing what the MMT preaches, just in a slightly more thoughtful form than all the Spend-Until-You-Die politicians promoting MMT are suggesting. Hence people should be excused for attacking AOC's MMT, rather than the scholarly MMT.

    ReplyDelete
    Replies
    1. The reality is that various “senior” MMT academics have ties to the Democratic Party; Stephanie Kelton had some big budgeting post or another. (I’m Canadian, and try to avoid being drowned in news about US politics, so I skip the details.) With that background, there’s no surprise that there are ties between that narrow group and the Democrats. But if we look at the attendees of the MMT conferences, it’s more of a mix, although I am pretty sure the average leans left.

      I’m Canadian, and have single-payer, highly subsidised universities, subsidised child care, etc. - and a 50% marginal tax rate. Although I like to point out that capitalists tend to blow up capitalism, I’m still a fan of capitalism as it exists here. I’m not too far from an Eisenhower Republican (from what I know about that extinct species), so I’m not exactly a “leftist”. The thing is, the Republican veered towards John Birch Society thinking (who thought Ike was a commie), and if someone believes that only gold is money, they won’t fit in with MMT. In other words, MMT is compatible with “small government conservatism,” but the problem is that the actual small government people in the real world are drifting into crypto-Austrianism that is incompatible with MMT.

      The Green New Deal is awkward from my perspective. It only makes sense if you believe the scientists who are worried about climate change. However, that belief seems to be a political litmus test. If someone agrees with the GND objective (de-carbonise), you’d want to listen to MMTists than mainstreamers - who are advocating a policy that would almost certainly fail (a carbon tax). (A upward-ratcheting carbon tax might have achieved the objective if implemented 20 years ago, not now.) But if one disagrees with the objective, that’s not MMT’s problem. This is my concern: by equating the GND to MMT, you end up subjecting MMT to the political litmus test that is climate change belief.

      Now back to AOC. She obviously ignited an interest in MMT, and that really helped my book sales, so I owe her a drink (if I ever met her). But one could just as easily use MMT to justify the Trump tax cuts. So I’d argue that if people were sensible, they’d argue policies on their merits - do we want this to happen? - using MMT as the analysis framework. If you think she’s out to lunch, just explain what’s wrong with her policies.

      Delete
    2. (same guy here). Her policies are, imho, misguided because they are overly dependent on MMT working miracles. The truth is, I believe, that an MMT gov/CB will lead to similar (or maybe more inflationary) budgets and monetary expansion, /unless/ they overestimate the possibilities of MMT, in which case it'll lead to a Weimaresque disaster.

      Even if (and I am no authority to judge) MMT is correct, it is a reinterpretation of market forces, not a new set of fundamental rules.

      Delete
    3. If you are referring to AOC, the only proposal of hers that I looked at was the GND. The actual proposal: set up a committee to develop a plan. Other than the hot air of the Congressional committee making climate change worse, there is no measurable effect on the economy. Need to wait for the committee to present proposals before even thinking about analysing it.

      Delete
  5. I have express a few of the articles on your website now, and I really like your style of blogging. I added it to my favorite’s blog site list and will be checking back soon… internships

    ReplyDelete
  6. Learning is a one of the best skill. Interesting student never miss this topics. Typical Student gives opportunity to students for better career .

    ReplyDelete
  7. This comment has been removed by a blog administrator.

    ReplyDelete
  8. This comment has been removed by a blog administrator.

    ReplyDelete

Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.