The Return of MMT at the ANU

I recently saw a very interesting debate over the merits of Modern Monetary Theory and Post-Keynesian ideas amongst some economics students at the ANU. The Modern Monetary Theorist (MMTer) claims can be summarised as follows:

Government taxation need only be driven by the need to force convertibility of the currency. Past that point, taxation as a means of revenue sourcing is unnecessary since the central bank can just monetise debt. Given that under MMT models, monetary policy is just an exchange of financial assets, such monetisation would not have an inflationary impact as long as the economy stays just below the NAIRU. Furthermore, if there were a small inflationary impact resulting from such policy, the evidence suggests that unemployment is a far greater burden on society than inflation.

I will not go over the basic principles of MMT. Such resources can be found here. Rather, I wish to respond more broadly to the debate at hand, without delving too much into the specifics.

As I perceive it, the crux of this argument stems from a misunderstanding of monetary theory that ignores the role of expectations and confuses the definitions of credit and money. My aim is to illustrate some of the nuance lacking in the argument above, whilst also confirming that once we look at the situation from the framework of modern macroeconomics the inflationary impacts of such an MMT policy proposal would be far more severe than alleged.

Firstly, the idea that government debt can be monetised without inflationary impact is one that can be answered by monetary theory alone. When we discuss the monetisation of deficits we are talking uniquely about the creation of base money in exchange for Treasury bills that would otherwise be sold to the market. Thus, the tangential idea that government spending cannot be inflationary since any government deficit must imply a foreign/private surplus is not relevant. This common retort, simply uses real demand side analysis to peer into the world of nominal relationships, and in facts relies on the assumption that the deficit itself is not being funded by seigniorage.

In this light, one does not need to talk about the monetisation of deficits in some isolated theoretical vacuum; rather we just need to look at the conduct of monetary policy more generally and in particular the theory underlining Quantitative Easing (QE). As I see it, the monetisation of a government deficit over time is the effective equivalent to a bout of QE, with the added proviso that the central bank commits to never unwinding the purchase.

With this clarification comes what Scott Sumner coined as ‘The Achilles Heel of MMT’. Let’s assume that the central bank opted to monetise a deficit via an expansion of its balance sheet that would never be unwound. As an example, let’s say they doubled the stock of base money. Neither banks nor the public are going to hold twice as much base money at the same interest rate, the opportunity cost is too high. Additionally, anticipating that such injections are permanent with other agents rebalancing their portfolios to clear off excess liquidity one would expect a flow of funds into asset and debt markets. This would irrefutably drive interest rates down to zero and inflate asset prices across the economy. Hence, we are left with an economy far outside its Wicksellian equilibrium. Continuing in this vein Sumner has noted:

 “MMTers forget that the nominal interest rate is the price of credit, not money.  The Fed can’t determine that rate, it reflects the forces of saving supply and investment demand.  Hence an attempt to set interest rates far below their correct level in savings/investment terms (the Wicksellian natural rate), would trigger an explosion in AD, and much higher inflation.  Central banks know this, which is why after the inflationary 1965-1981 period they adopted the Taylor Principle.”

In response to this argument, many MMTers would cite cases of Japan post 1990 and the USA in 2009, situations where the economy was far below the NAIRU. But this is a dishonest citation with an insincere framework. Before these case studies can be addressed, two fundamental distinctions have to be made. Firstly, the process of expanding central bank balance sheets is one that is primarily concerns bond, not labour, markets. Thus, the NAIRU framework for looking at monetary operations is not an accurate one. Rather, we need to look at these cases through the lens of a good ol’ fashioned liquidity trap. Whilst changing this criteria may seem somewhat trivial, when it comes to the distortion of facts it is an important distinction. MMTers assert far too often that since an economy is almost never at the NAIRU inflationary pressures are not a concern. From the perspective of monetary policy nevertheless, inflationary concerns are always present when expanding the balance sheet unless the economy is in a liquidity trap. Liquidity traps however, are a far rarer observation than economies with employment below the NAIRU. Thus, MMTers claiming that balance sheet expansions are anywhere and everywhere inflation neutral are either insane or dishonest.

The second distinction however is even more important. It is true that in both Japan’s lost decade and the US’ post-GFC stagnation central banks greatly expanded their balance sheets, increasing the supply of base money by trillions of dollars. Nevertheless, in both these situations the QE operations were perceived by the market to be temporary operations. The distinction between temporary central bank expansions of the balance sheet and a permanent increase via the monetisation of deficits is an important one. Let us consider an example.

In the first case, consider a financial institution temporarily credited by the central bank with base money. Optimally, it wants to profit off of this credit, and as such it may purchase some financial assets. Given that this operation is temporary though, at some point this purchase must be unwound with a corresponding sale. As a group, all financial institutions will have to unwind purchases with sales at some point. With arbitrage accordingly the fundamental demand for assets over time will remain unchanged.

Now consider a financial institution that is permanently buffeted by central bank liquidity injections. In this case, the purchase of an asset doesn’t have to be unwound at any point, so any current purchases need not be met by future sales. Here, there is no scope for arbitrage and asset markets witness a tangible increase in demand.

Consequently, the difference between a temporary and permanent expansion of central bank balance sheets is immense. Since monetisation of fiscal deficits must imply a permanent iteration of QE the effects would diverge considerably from those observed in Japan and the US.

At this point we are left with just a reiteration of the standard rebuttal that “unemployment is worse than inflation, so even if MMT policy proposals would result in higher inflation, the benefits of reducing unemployment would far out-weigh the costs.” At this point we are beating a dead horse. The notion that there is a long-run trade-off between inflation and unemployment is one that has been repeatedly rejected by both theory and empirics since the 1950s. In the short run, I am sure that MMT proposals could boost output, especially in Europe where a credible commitment to increase the money supply is desperately needed. But in the long run, such ideas are of no merit whatsoever.

It is for these reasons that I see the position of MMTers as untenable when it comes to deficit monetisation. The proposals put forward are invariably inconsistent with the current monetary theory and offer no alternative framework of their own.

3 comments

  1. That $10 injection with a $20 payoff of course would also correspond to the case of a valuable ‘public good’. The point as I see it is divorced from the real side of the economy and is all about monetary policy

  2. Thanks for a great analysis! The one concern I have still is that it seems to me as though an MMT advocate (certainly the one you refer to on facebook) would say that government can offset any increase in the money supply by spending on things that create a requisite increase in production and thereby prevents inflationary pressure. I’m not quite sure what to make of this response. I would infer that this would lead to an expenditure model where government taxes to fund consumption goods (e.g. art, welfare), but not capital goods (education, infrastructure). I’m interested in what you make of this response. It occurs to me that you probably just end up with a crowding out effect, but the logic of public goods would suggest to me that the public sector could probably find projects for which it could match the private sector on economy-wide returns.

    1. Hi Tom,

      That’s a good question, so I hope I can answer it. If I haven’t, just let me know and maybe the MMTers will take the day ;). When you say that “government can offset any increase in the money supply by spending on things that create a requisite increase in production” there are a few things going on and as you suggest, crowding out plays a big role.

      To thoroughly answer this question I think it is important to distinguish between ‘financial’, and ‘resource’ crowding out and to also make two possible hypotheses about how the economy operates- then take the actions to their logical extreme.

      This plays into the NAIRU v Liquidity Trap framework. Indeed, I don’t think it’s sufficient for MMTers to state “an increase in production will offset the inflationary pressure” in all circumstances where we lie below the NAIRU. Rather, their ideas require that we are in a far more severe ‘liquidity trap’ where base money is hoarded by the banking sector.

      For the first hypothesis, let’s assume that there is no slack in the economy. Clearly, government purchases will crowd out private spending, and the increased money supply will have the obvious inflationary effect.
      For the second hypothesis, let’s assume that there IS slack in the economy, and that the government can create real production. (This is the more interesting case).

      Whilst it is true that it is possible for the government to create real production when the economy is in this situation, it is not true that permanently expanding the money supply alongside it will be inflation neutral. Consider the case where say, a government injection of $10 creates a real production boost of $20 of GDP, but where the deficit is monetised by the central bank. For the real side of the economy, this is great, we get a $20 increase in output but look at how monetary policy needs to respond. Given that we’ve set up a ‘QEesque’ framework for understanding deficit monetisation, I claim that this government spending must imply a $10 increase in the level of base money. With this, we know that even though we’re below the NAIRU, the banking sector may still increase the money supply by up to $100! Thus, if lending growth is still healthy we have huge inflation risks. As a result, for our monetary experiment to be inflation neutral we need to impose stricter conditions. Namely, we require some sort of liquidity trap whereby reserves get stockpiled in the system and do not translate broadly into money. Liquidity traps are however very rare and not suitable conditions to generalise economic policy into normal times.

      In fact, I would extend this further and claim that liquidity traps are actually impossible if we are monetising deficits. This is because the liquidity injections here are by definition permanent, whilst in historical liquidity traps all monetary responses have been temporary in nature.
      Note here, that the difference between deficit spending WITH monetisation and deficit spending WITHOUT is that the latter incorporates financial crowding out, whilst the former does not. Thus, even though in both cases there is no ‘resource’ crowding out, we still need to pay attention to phenomena in financial markets. The inflationary bias of MMT deficit monetisation comes from the fact that we remove the ‘financial’ crowding out which otherwise operates normally.

Leave a Reply