A Comment on Lavoie’s MMT Paper

In my previous post, I linked to an interesting working paper by Marc Lavoie in which he takes a friendly critical look at MMT. While considering MMT analysis to be essentially correct, and pointing out significant areas of consistency with Post Keynesian perspectives, particularly of the circuitist and horizontalist persuasions, he questions the need for some of the more controversial aspects of the approach. His central focus is on the MMT abstraction of collapsing the Treasury and central bank (referred to as the Fed in the case of the U.S.) into the consolidated government sector.

In his conclusion, Lavoie writes:

There is nothing or very little to be gained in arguing that government can spend by simply crediting a bank account; that government expenditures must precede tax collection; that the creation of high powered money requires government deficits in the long run; that central bank advances can be assimilated to a government expenditure; or that taxes and issues of securities do not finance government expenditures. All these counter-intuitive claims are mostly based on a logic that relies on the consolidation of the financial activities of the government with the operations of the central bank, thus modifying standard terminology. I believe that such a consolidation leads to the avoidance of crucial steps in the analysis of the nexus between the government activities and the clearing and settlement system to which the central bank partakes, and hence leads to confusion and misunderstandings.

In my view, if the abstraction of the consolidated government sector is an appropriate one, MMTers would be correct to state explicitly any conclusions that follow from it, even if these might seem counter-intuitive or controversial. The more important question is whether the abstraction is appropriate.

My current understanding is that the abstraction is a good one to make, but I will be interested to see the responses of the academic MMTers. A good resource on the topic is provided by Fullwiler in his introductory post on monetary operations.

Even if MMT analysis of the consolidated government sector is appropriate, at the very least Lavoie’s discussion is illuminating in identifying key points in dispute.

Consider this passage from p.18:

What seems to truly happen in the USA is thus illustrated by Table 3 below, which reproduces in T-accounts the sequence most recently described by Wray in the same blog: “So, instead, the Treasury sells the treasuries to the private banks, which create deposits for the Treasury that it can then move over to its deposits at the Fed. And then ‘Helicopter Ben’ buys treasuries from the private banks…. The Fed ends up with the treasuries, and the Treasury ends up with the demand deposits in its account at the Fed – which is what it wanted all along, but is prohibited from doing directly” (Wray 2011C). In the first step, as in Table 2, the government sells its securities to the commercial banks. In the second step, the government deposits are shifted from the commercial banks to the central bank, thus creating a negative reserve position for banks. The central bank then takes defensive compensatory measures, purchasing back the Treasury bills on the secondary markets, and thus eliminating the deficiency in bank reserves at the Fed.

So, due to current self-imposed arrangements, in step 1 the Treasury, rather than selling treasuries directly to the Fed, first sells them to private banks. The private banks pay for these treasuries (assets of the private banks) by crediting the government’s accounts with the private banks (liabilities of the private banks).

In step 2, the government transfers its deposits from the private banks to the Fed. For the private banks, there is a decrease in government deposits (private bank liabilities) and a debiting of reserve accounts (private bank assets). For the government, there is an increase in deposits with the Fed and a matching decrease in deposits with the private banks.

The combined effect of steps 1 and 2 on the private banks is to reduce their reserves but increase by a matching amount their holdings of treasuries, leaving private bank assets unchanged.

To replenish reserves, in step 3 the Fed purchases the treasuries from the private banks by crediting reserve accounts. Again, this leaves private bank assets unchanged.

The private banks are back to the point they were before the beginning of the three steps. In aggregate, they temporally held and then relinquished treasuries, they credited and then debited the government’s accounts, and they lost and then regained reserves.

In contrast, in the government sector, the Fed has extra treasuries and the Treasury has extra deposits in its account at the Fed.

The overall effect of these three steps is as if the Treasury had simply sold treasuries directly to the Fed, but this is prohibited under current arrangements, so the same effect is achieved in a roundabout way.

Okay, so far everyone agrees, Lavoie and the MMTers.

On pp.18-19, Lavoie continues:

The purpose of this whole exercise is to show that there is no point in making the counter-intuitive claim that securities and taxes do not finance the expenditures of central governments with a sovereign currency. Even in the case of the US federal government, securities need to be issued when the government deficit-spends, and these securities initially need to be purchased by the private financial sector. It seems to me that the consolidation argument – the consolidation of the central bank with the government – cannot counter the fact that the US government needs to borrow from the private sector under existing rules. Thus, if even the USA does not really fit the bill, one may wonder whether there is any other nation that corresponds to the strictures of neo-chartalism.

It is clear that the government only needs to issue treasuries because of the self-imposed arrangement that is in place. It would actually be much simpler to do away with treasuries altogether and simply spend as required and allow reserves to mount as a result, and then pay the interest rate on reserves. By stating this point clearly, MMT makes transparent that other, simpler arrangements could achieve the same effect.

However, since this is not currently allowed, it raises the issue – as Lavoie discusses – of whether government deficit spending should be considered as funded by the private sector.

To consider the nature of steps 1-3, it may help to ponder a series of questions and answers. In step 1, the Treasury sells securities to the private banks. What do the private banks do in exchange for the treasuries they receive? They create private bank deposits for the government. Why is this an acceptable exchange from the perspective of the government? Simply because the government deems the created private bank accounts (bank money) to be transferrable into reserves (government money), which it is willing to accept in fulfilment of the tax obligation it imposes on the non-government. Why are private banks satisfied? Precisely because the treasuries are convertible into reserves, and these are what are needed to extinguish tax obligations.

In step 2, the Treasury moves its account to the Fed and the Fed debits the reserve accounts of the private banks. Why is this acceptable to the private banks? Because they know that in step 3 they can get the reserves back again in exchange for the treasuries they still hold. Do the private banks lose anything in this process? No. At the end of the three steps there is no change in their holdings of treasuries, their reserve levels or deposit liabilities.

In short, there is no risk of this convoluted process breaking down. The Fed can always create reserves in exchange for the treasuries issued by the Treasury, and will always do so to the extent that this is necessary to enable banks to meet reserve requirements. The Fed is lender of last resort. Any action of the Treasury that causes an excess or deficiency of reserves must be offset by the Fed. By necessity, the Fed works in tandem with the Treasury.

The Treasury and Fed, in their actions, behave as two departments of the government sector.

The MMT position is therefore that, for most purposes, it makes sense to abstract from the details of the convoluted three-step process and collapse the Treasury and Fed into the consolidated government sector. These details add no additional insights except when the object of analysis is these details themselves.

Once the Fed has the securities and the Treasury has the matching deposits at the Fed, it is straightforward to understand that government spending leads to a net increase in non-government bank accounts and a corresponding rise in reserves, which are then drained, if there is a positive interest target, through further purchases of securities.

It is only at this point – when the government spends – that the non-government’s net financial assets are affected. For this reason, most elementary MMT renditions of fiscal policy commence at the point where the government spends and taxes by crediting and debiting non-government bank accounts. Then, to the extent spending exceeds taxes, the Fed will normally drain excess reserves by purchasing securities (this is not necessary when the short-term interest rate is zero).

This is why it seems appropriate to think of government spending first and draining reserves at the end of the process. It also makes sense to state that government spending (state money creation) is logically prior to taxing (state money destruction). Logically, the capacity of the non-government to pay taxes or purchase bonds comes after the government has spent or the Fed has lent, because the payment of taxes or the purchase of bonds requires the prior creation of reserves.

This point is clearly expressed by Fullwiler in the post linked to above:

This all leads me to the often noted MMT point that “spending comes before tax revenues are received or bond sales.” If one expands this a bit to include loans from the Fed, then this statement is absolutely correct in terms of the operational realities of the monetary system. That is, according to both the tactical and accounting logics, taxes credited to the Treasury’s account and the settlement of Treasury bond auctions can only occur via bank reserve accounts, while the original source of banks’ balances in their reserve accounts can only be previous government deficits (which are net credits reserve accounts) or loans from the Fed (repos, loans, purchases of private securities, or overdrafts—note that an outright purchase of a Treasury security by the Fed to add reserve balances requires a previous government deficit). Therefore, it very much is the operational reality that for taxes to be paid or bonds to be settled, there has to have been previous government spending or loans from the Fed to the non-government sector, and this is true whether or not the Fed is legally prohibited from providing overdrafts.

With this point in mind, it is worth reflecting one more time on the convoluted three-step process between the Treasury, Fed and private banks that the government imposes on itself. Because this process ultimately has no aggregate effect on reserves, it is clear that it is not the place to look when asking where the funds to pay taxes originate. The capacity of the non-government to pay taxes comes from reserves, and these are always the result either of prior government spending or Fed lending.


103 thoughts on “A Comment on Lavoie’s MMT Paper

  1. I have to play Devil’s Advocate here…

    “In step 2, the Treasury moves its account to the Fed and the Fed debits the reserve accounts of the private banks. Why is this acceptable to the private banks? Because they know that in step 3 they can get the reserves back again in exchange for the treasuries they still hold. Do the private banks lose anything in this process? No. At the end of the three steps there is no change in their holdings of treasuries, their reserve levels or deposit liabilities.”

    Certainly, the private banks will eventually be provided with the reserves they need to meet reserve requirements after any loans that they desire to make have taken place.

    But we are not talking about a purely static situation here. Time has to intervene. And, if my understanding of these things is correct, the private bank may, in the interim between the holding of treasuries and the next Fed intervention in the interest rate, make enough loans that they have to hit the discount window. This would eat into their profits — and we can be sure that, good capitalists that they are, they’re keenly aware of these.

    If I’m right on this (and I may not be, by all means correct me if I’m wrong) the MMT approach of aggregating the sectors may be veneering over some ‘friction’ in the system — friction that does indeed determine private banks’ actions under certain conditions.

    (In fact, whether I’m right or wrong on this I should say now that I’ve long thought that the MMT literature on the discount window and when it is used etc. was a tad vague. Maybe someone could point me in the right direction?)

Leave a Reply

Your email address will not be published. Required fields are marked *