Functional Finance Versus New Keynesian Economics, Krugman Edition

Paul Krugman has piled onto the "MMT explained by non-MMTers" bandwagon, with a critique of Functional Finance. Functional Finance is largely associated with the Old Keynesian Abba Lerner, and is one of the key intellectual roots of Modern Monetary Theory (MMT). In my view, the most interesting part of the article is that it contradicts the commonly made assertion that there is very little new in MMT (which Krugman hints at in the article as well). In presenting his summary of Functional Finance, Krugman obviously has theoretical blinders on, and the objective of MMTers is to point out the existence of those blinders.

There is a legitimate substantive debate about issues underneath the disagreement, I am not going to assert which side is correct. However,I would note that only the MMT side actually sounds like it made the effort to understand the ideas on both sides of the debate. As a result, there is no doubt that MMT represents an advancement of knowledge relative to the neoclassical consensus.

For further reading, I wrote a primer on Functional Finance (which I worked into Understanding Government Finance).  As a disclaimer, I wrote that primer to summarise the original scholarly article by Lerner, and I did not attempt to capture the evolution of Lerner's thought. Krugman's discussion of Functional Finance appears to be based on the more fleshed out version that was developed later. This distinction probably needs to be kept in mind: although the principles of Functional Finance were incorporated into MMT, this does not mean that everything that Lerner wrote is aligned with the MMT view. As such, there is little value into delving into the details of Lerner's views if the objective is to discuss MMT, rather what the MMTers actually use.

Krugman correctly expected a response similar to mine:

Unfortunately, that’s a very hard argument to have – modern MMTers are messianic in their claims to have proved even conventional Keynesianism wrong*, tend to be unclear about what exactly their differences with conventional views are, and also have a strong habit of dismissing out of hand any attempt to make sense of what they’re saying. The good news is that MMT seems to be pretty much the same thing as Abba Lerner’s “functional finance” doctrine from 1943. And Lerner was admirably clear, making it easy to see both the important virtues of and the problems with his argument.

It would not be a major stretch to call MMTers scrappy when it comes to online debate. But at the same time, Krugman is hardly coming off as non-combative in that paragraph either. If someone who obviously does not understand your world view mangles it and then explains why the mangled version is incorrect, one is not expected to play along.

As for "what exactly their differences with conventional views are", Krugman manages to demonstrate that himself, as I discuss below.

Krugman argues there are two main issues with Functional Finance.

1. Functional Finance does not take monetary policy seriously enough.
2. The r>g condition.

I discuss these in turn.

From a modern perspective, “Functional finance” is really cavalier in its discussion of monetary policy. Lerner says that the interest rate should be set at the level that produces “the most desirable level of investment,” and that fiscal policy should then be chosen to achieve full employment given that interest rate. What is the optimal interest rate? He doesn’t say – maybe because through the 30s the zero lower bound made that point moot.
Anyway, what actually happens at least much of the time – although, crucially, not when we’re at the zero lower bound – is more or less the opposite: political tradeoffs determine taxes and spending, and monetary policy adjusts the interest rate to achieve full employment without inflation. Under those conditions budget deficits do crowd out private spending, because tax cuts or spending increases will lead to higher interest rates. And this means that there is no uniquely determined correct level of deficit spending; it’s a choice that depends on how you value the tradeoff.

The whole premise of Functional Finance is that inflation control was largely a question of fiscal policy. Who cares if the approach to monetary policy is "cavalier" if the modern perspective is cavalier in its dismissal of fiscal policy (except for the ZLB exception, which is brought up continuously)?

The Functional Finance view is only a problem if it can be demonstrated that interest rate policy is superior to fiscal policy in management of the economy. The mainstream has convinced itself that this is the case, and have argued fervently for decades in favour of the supremacy of monetary policy. However, they have not convinced everyone, with the MMTers notably holding contrary views. This is a glaring difference in world view. Given that rather impressive divide, how is it even slightly possible that Krugman can argue that MMTers "tend to be unclear about what exactly their differences with conventional views are"?

As for the technical part of Krugman's comments, I am somewhat mystified as to their significance. By assumption, if we are at "full employment" (which for New Keynesians is normally some variant of NAIRU, the level of which nobody can really pin down in real time), by definition, we cannot add jobs (without inflation, anyway). So obviously, increasing the deficit does nothing. So what? Even under Functional Finance, if we accept the premise of "full employment," you get exactly the same result. However, as MMTers point out, in the current system, we control inflation by throwing a portion of the population into unemployment. Fiscal policy -- such as a Job Guarantee -- humanely deals with the realities of capitalism by directly providing income to those without a private sector job. Furthermore, income support (stimulus) is directed at regions that are weaker economically; geographical targeting is impossible with monetary policy. Furthermore, the reality is that monetary policy largely "works" in recent decades via creating housing bubbles, which is a trade-off that Krugman ignores.

We then get to the r>g question.

What about debt? A lot depends on whether the interest rate is higher or lower than the economy’s sustainable growth rate. If r<g, which is true now and has mostly been true in the past, the level of debt really isn’t too much of an issue. But if r>g you do have the possibility of a debt snowball: the higher the ratio of debt to GDP the faster, other things equal, that ratio will grow. And debt can’t go to infinity – it can’t exceed total wealth, and in fact as debt gets ever higher people will demand ever-increasing returns to hold it. So at some point the government would be forced to run large enough primary (non-interest) surpluses to limit debt growth.

Alexander Douglas wrote an article on this aspect of the Krugman piece. Douglas' response:

Like many other recent criticisms of Functional Finance, Krugman’s criticism reduces to the question: ‘what if r>g?’. But the burden doesn’t lie on the defender of a policy to explain what happens if it isn’t implemented. The idea is to implement it.

Douglas' argument is premised on the idea that the interest rate (r) is a policy variable. According to the neoclassical view, the interest rate has to revert to some variant of the natural rate of interest. From the heterodox perspective, there is no natural rate of interest, and so the level of r is a choice. Krugman's text is premised that the r>g relationship is some form of a law of nature, such as a gravitational constant. This is not true from a MMT/Functional Finance perspective. In particular, if we move to the MMT version of Functional Finance, we can dispense with this by locking the nominal rate at zero. All the government needs to do is force nominal GDP to grow faster than 0%, then r < g. So for the case of MMT, the r>g criticism makes no sense.

What was that again about there not being clear theoretical differences between the conventional view and MMT?

(I will return to the "debt snowball" with in my Technical Appendix.)

Concluding Remarks

Although I love pointless theoretical debates as much as the next econ blogger, my view has been that the best strategy is to focus on substantive issues. However, the deluge of bad MMT takes has forced my hand, and I have been dragged back into economic squabbling...

Technical Rant Appendix

The following passage caught my eye, and I was so triggered, that I needed to respond. Krugman:

And debt [BR: from context, this refers to the debt-to-GDP ratio] can’t go to infinity – it can’t exceed total wealth, and in fact as debt gets ever higher people will demand ever-increasing returns to hold it. 

The part about [debt] "can't exceed total wealth" makes no sense, so I will assume it is some form of a typo. (It makes no sense since government debt is part of private sector wealth, and so increasing the stock of government debt increases total wealth.) I will instead comment about debt-to-GDP ratio going to infinity. (The stock of debt is unbounded in any model with nominal growth; from context, we are worried about the debt ratio.)

My claim is that in any plausible economic model, it is impossible for the ratio of government debt to GDP to go to infinity. The implication is straightforward: the only way for the debt-to-GDP ratio to "go to infinity" is in the context of a implausible economic model. So why bring up the possibility in the first place?

Why cannot the debt/GDP ratio go to infinity? Unless there are shenanigans in the form of unbounded circular flows of lending between the government and the private sector (e.g., the government lending money to individuals so that they can buy government bonds), government debt holdings will translate into net wealth for at least some individuals. The greater the stock of debt, the larger the wealth. If the debt-to-GDP ratio becomes unbounded, then the ratio of the wealth of particular debt holders to GDP has to become unbounded (under the mild assumption that the Earth has a finite maximum population).

We would end up in a situation where an individual could buy 100% of a nation's output with just 0.00000001% of their financial holdings. Such a situation seems implausible, to put it mildly; the individual would just buy everything up. If expectations matter, everyone else would have seen this coming, and raised the price level (wiping out the debt-to-GDP ratio via inflation).

In other words, in any sensible economic model, the debt-to-GDP ratio cannot march to infinity. And if we look at stock-flow-consistent models, outcomes meet that requirement for plausibility.

The issue the mainstream faces is that despite claims of mathematical rigor, nobody bothers to actually calculate model trajectories. They just assume that the steady state values of r and g are fixed, without actually seeing what happens if that is the case. The reality is that if the models truly say that the debt-to-GDP ratio is going to infinity, the model dynamics imply some laughably bizarre outcomes.

Say what you want about MMTers, they do not kill trees and/or electrons opining about scenarios that will obviously never happen.


Footnote:

* One may note the sloppiness in terminology: what is "conventional Keynesianism"? Although I argue that the obsession with Keynes is perhaps not the strongest point of post-Keynesianism, they have mapped out admirably well the variants of Keynesianism: old school Keynesians (like Lerner), the neoclassical New Keynesianism (that are effectively Monetarists with sticky prices) and post-Keynesianism. MMT academics unabashedly describe themselves as post-Keynesians, so it is abundantly clear where they fit into the "Keynesian" spectrum.


(c) Brian Romanchuk 2019