Paying interest on excess reserves is not constrained by scarcity

This morning, a former deputy governor of Australia’s central bank (RBA) published a short Op Ed in the Australian Financial Review (July 16, 2019) – Why there are no free lunches from the RBA – which served as a veiled critique of Modern Monetary Theory (MMT). The problem is that the substantive analysis supported the core of the MMT literature that we have developed over 25 years, refuted the standard macroeconomics textbook treatment of the link between the government and non-government sectors, and, incorrectly depicted what MMT is about – all in one short article. Not a bad effort I thought. But disappointing that a person with such experience and knowledge resorts to perpetuating such crude representations of ‘cost’ and myths about government finances.

The author, Stephen Grenville, now a consultant to the World Bank, IMF and other like institutions, seeks to caution against the “current vogue for asking the central bank to fund socially desirable expenditures”.

He is somewhat mistaken to source the start of this ‘vogue’ as “America’s quantitative easing (QE)” which started in 2008. Japan has been demonstrating what central banks can do in this regard when they adopted QE on March 19, 2001.

At that point, the Bank of Japan pushed large volumes of ‘excess reserves’ into the commercial banking system in return for purchases of government and by ensuring there were always excess reserves each day in the cash system they have been able to maintain short-term interest rates at zero ever since.

It is clear that in response to the GFC, other central banks – the Federal Reserve in the US, the Bank of England, Swiss National Bank, the Sveriges Riksbank, and the ECB – followed suit in varying proportions.

The motivations were expressed differently but the intent was to militate against the recessionary impacts of declines in non-government spending.

What was held out to the public was a policy approach to push more reserves into the system to enhance the ability of the commercial banks to make loans at times when credit was tight and economic activity was in decline.

Modern Monetary Theory (MMT) economists, at each phase, pointed out that this justification was based on the false mainstream notion that banks loan out reserves and that lending can become reserve constrained.

The reality is that banks only loan excess reserves among themselves as part of the payments system (cheque clearing) and that lending is not constrained by deposits (and hence, reserves).

MMT economists were the first in the modern era to point out that loans create deposits not the other way around. You will never find that proposition in the standard macroeconomics textbooks.

What follows is that QE cannot influence bank lending in the way claimed by mainstream economists. The lack of lending behaviour during the GFC was the result of a shortage of credit-worthy borrowers.

People were simply reluctant to borrow given elevated and precarious debt levels and this constrained the loan books of the banks.

The only way QE might have been a stimulative measure was in its impact on interest rates at the ‘investment’ end of the yield curve. By purchasing the volumes of government debt that QE achieved, the central banks drove up bond prices (via higher demand) and pushed down yields.

Competition in the financial asset classes at the maturities of the bonds being bought then led to lower interest rates across the board.

But, as we learned, in times of heightened economic uncertainty, people will not borrow even as the cost of funds fall, if they lack confidence in the returns that the projects might deliver into the future.

Stephen Grenville claims that QE is “printing money” and that:

The revival of Modern Monetary Theory (MMT) promotes the same idea.

First, I am not sure the term “revival” is very descriptive. It suggests that MMT had a day in the sun, faded and now is shining again.

I think it is more accurate to think of MMT as a growing body of work which people are finding out about slowly but surely and realising that it helps them understand the dissonance between what mainstream economists are saying and predicting and what MMT says.

Second, to characterise MMT as advocating “printing money” is, of course, incorrect and demonstrates a rather superficial knowledge of the core literature.

The term is also loaded – and critics know that and use it intentionally. It invokes years of characterisations of out of control central bankers goaded by government officials running printing presses at ever increasing speeds and people wandering the streets carrying wheelbarrows full of bank notes to buy a loaf of bread.

Stephen Grenville adds some weight to his accusation, claiming that (MMT says):

It appears that budgets can be funded costlessly by printing money because money doesn’t pay interest. Sadly, this seemingly costless funding source isn’t, in fact, a “free lunch”.

I will come back to this claim after going through the logic of his argument.

That logic actually correctly depicts the way the central bank interacts with the commercial banks and the impacts of government deficits on the banking system.

In that sense, it captures what MMT economists have been writing about for over 25 years. It also exposes the lack of insights in the mainstream macroeconomics.

I believe our new MMT textbook – Macroeconomics – is the only one on the market that actually gets this all down correctly.

So Stephen Grenville is, in fact, while trying to critique MMT, unwittingly defending core components of our analysis.

And, unintentionally providing a sound critique of mainstream money and banking analysis.

He considers two scenarios.

First:

Suppose the central bank funds infrastructure expenditures, receiving in return some kind of zero-interest government IOU.

Note, he thinks of the central bank as being separate from the government, when in fact, central banks are legal and politicial creatures of the government and have to work daily in close cooperation with the treasury arm of government in their liquidity management functions.

But lets play along with the story.

In other words, the government tells the central bank to pay some bills to, say, road building companies etc and offers them a bit of paper in return bearing no interest.

The central bank credits the bank accounts of the infrastructure providers or in Stephen Grenville’s words “creates central bank money”.

There is no money multiplier operative as in the mainstream textbooks. Stephen Grenville’s version of this is that “the banks are already lending to all the borrowers who are judged to be bankable”.

Which is another way of saying that loans create deposits not the other way around (the latter being required for the operations of the fictional money multiplier) and if banks are lending to all the credit worthy borrowers requesting credit then they will not create any further deposits.

So what fiscal deficits do in this case is expand excess reserves – after all the transactions are exhausted – the liquidity injection from the net public spending shows up as excess bank reserves – these are held in accounts the commercial banks have to hold with the central bank.

I detail that process in detail in this introductory suite of blog posts:

1. Deficit spending 101 – Part 1 (February 21, 2009).

2. Deficit spending 101 – Part 2 (February 23, 2009).

3. Deficit spending 101 – Part 3 (March 2, 2009).

In Stephen Grenville’s words:

… the bank deposits the excess funds with the central bank, in the form of bank reserves.

Pure MMT.

So what is his point?

Well because:

Central banks just about everywhere pay a market-related rate of interest on reserves … Thus the expenditure is funded by the banking system holding excess reserves, on which the central bank pays a market-related interest. No free lunch here – just a distortion of the banks’ balance sheets, holding more reserves than they need. It would be better if this expenditure were to be funded, in the conventional way, by the government issuing bonds to the public.

First, whether central banks pay a “market-related rate of interest on reserves” is a policy choice.

The RBA pays “an interest rate on ES balances that is 0.25 percentage points below the cash rate target” (Source).

ES balances are the ‘exchange settlement’ accounts (reserves) that banks are reequired to maintain with the central bank.

Since the GFC, the US Federal Reserve pays an IOER (interest on excess reserves) – currently set at 2.35 per cent (as at February 5, 2019). This is in line with the “the target range for the federal funds rate at 2-1/4 to 2-1/2 percent” (Source)

The Bank of Japan has a different policy approach to excess reserves. It deploys a three-tier approach – “to which a positive interest rate, a zero interest rate, and a negative interest rate are applied, respectively. The three-tier system encourages negative interest rate transactions in the money market” (Source).

This exempts some proportion of the excess reserves from the negative rate applied at the margin. I may write about tiered reserve systems separately, given that the ECB is toying with the introduction along the Japanese lines.

The point is that it is a policy decision whether to leave excess reserves in the system to drive the short-term interest rates down to zero (as banks compete to offload their excesses onto other banks).

As Stephen Grenville acknowledges “The banks can swap these reserves among themselves, but the banking system as a whole can’t get rid of the excess central-bank money” (another pure MMT proposition that you won’t find articulated in mainstream textbooks).

And if there is no return offered, banks with excesses on a particular day will try to loan them to banks facing shortfalls. But if there is a system-wide surplus, this activity is fraught and only results in the yields dropping on the loans in the Interbank market to zero.

So if a central bank desires a non-zero policy rate then it has to either sell interest-bearing bonds to the banks to drain these reserves or just pay them a market rate for holding the excesses.

A policy choice.

Second, the concept of ‘cost’ that Stephen Grenville invokes lacks any sense. He seeks to exploit the concept of – There ain’t no such thing as a free lunch – which Milton Friedman’s 1975 book There is No Such Thing as a Free Lunch advanced as part of his attack on government activity.

But alleging that an interest payment from the central bank is somehow a ‘cost’ in the sense that Friedman and others use the term ‘no free lunch’ is invalid.

Friedman was talking about real resource scarcity and in an environment that sort of scarcity is a binding constraint there is an opportunity cost involved in making choices.

If we choose to use a resource in one way, then it cannot be used in another way. The choice thus influences the ‘net benefits’ that flow from that resource usage against the usage foregone.

Friedman and his clan always claimed that government usage of productive resources led to inferior outcomes because the ‘free market’ did not ‘discipline’ the choice.

So in economics, the ‘free lunch’ metaphor is about opportunity cost. At the heart of mainstream economics is the claim by, say, Gregory Mankiw that “To get one thing that we like, we usually have to give up another thing that we like. Making decisions requires trading off one goal against another.”

Of course, the concept is difficult to sustain at a macro level if there are idle resources – that is, productive resources that are being wasted as a result of unemployment.

Then, for a time, the opportunity costs are low to zero and there is a ‘free lunch’.

Now ask yourself, what the ‘next best alternative’ use of an interest payment made by the currency issuer is when that payment is accomplished by the stroke of a computer key and comes from ‘nowhere’!

In his famous paper – Fifteen Fatal Fallacies of Financial Fundamentalism (October 5, 1996) – American economist William Vickrey discussed said that:

Much of the conventional economic wisdom prevailing in financial circles, largely subscribed to as a basis for governmental policy, and widely accepted by the media and the public, is based on incomplete analysis, contrafactual assumptions, and false analogy.

He noted that these fallacies lead to “the assumption that future economic output is almost entirely determined by inexorable economic forces independently of government policy so that devoting more resources to one use inevitably detracts from availability for another.”

He also accepted that “This might be justifiable in an economy at chock-full employment …”.

But when economic policy is “keeping us in the economic doldrums with overall unemployment rates stuck in the 5 to 6 percent range … the loss of 10 to 15 percent of our potential production … unemployment of 10, 20, and 40 percent among disadvantaged groups, the further damages in terms of poverty, family breakup, school truancy and dropout, illegitimacy, drug use, and crime become serious indeed.”

In these situations, claiming there is ‘no free lunch’ is simply absurd ideological nonsense.

His 15 fallacies document beautifully what remains problematic with mainstream macroeconomics and the type of reasoning Stephen Grenfell tries to invoke in the article under discussion.

They relate to the standard arguments of ‘sound finance’ – you should know them all by now. They are the core propositions that MMT undermines.

William Vickrey wrote that:

These fallacious notions, which seem to be widely held in various forms by those close to the seats of economic power, are leading to policies that are not only cruel but unnecessary and even self-defeating in terms of their professed objectives … We will not get out of the economic doldrums as long as we continue to be governed by fallacious notions that are based on false analogies, one-sided analysis, and an implicit underlying counterfactual assumption of an inevitable level of unemployment.

And in conclusion he wrote with considerable eloquence that:

To assure against such a disaster and start on the road to real prosperity it is necessary to relinquish our unreasoned ideological obsession with reducing government deficits, recognize that it is the economy and not the government budget that needs balancing in terms of the demand for and supply of assets, and proceed to recycle attempted savings into the income stream at an adequate rate, so that they will not simply vanish in reduced income, sales, output and employment. There is too a free lunch out there, indeed a very substantial one. But it will require getting free from the dogmas of the apostles of austerity, most of whom would not share in the sacrifices they recommend for others.

The point is that a core principle of MMT is a recognition that what constrains government spending are the available real resources that can be purchased and put into productive use or diverted from other uses.

Whether there is scarcity or not depends on the state of utilisation of the resources available to a nation. That assessment then determines what policy interventions might be appropriate.

But there are no ‘real’ constraints on the central bank paying interest on excess reserves should they decide to do so. From an MMT perspective, there is no reason for the central bank to do so, just as there is no reason to match fiscal deficits with debt issuance.

Stephen Grenfell’s second scenario is, in his words “closer to the “printing money” idea: that the government can pay with interest-free liabilities.”

So, instead of crediting the bank accounts of the infrastructure supplier, he posits that the central bank loads up some delivery vans with actual bank notes and pays the supplier accordingly.

What happens then?

Well in his words:

The contractors don’t want the cash, nor do they want to splurge it on bidding up prices, as Milton Friedman warned. Instead, they deposit it with their bank. The bank doesn’t want the cash either, so it gives it back to the central bank in return for bank reserves. We’re back at the same place.

Which is pure MMT operational intelligence. You won’t find this sort of logic in any mainstream macroeconomics textbook.

Students in mainstream programs are taught the Milton Friedman line – that inflation would ensue.

At least Stephen Grenfell is experienced enough to understand that the inflation risk is embedded in the spending decision not how that spending is made operational.

His next logical statement is that “the central bank money ends up as bank reserves on which the central bank has to pay interest.”

Again, note that whether the central bank pays interest on excess reserves is a policy choice. It clearly doesn’t have to do so.

And the next link in his logic train fails similarly:

This interest payment may not show up as a line item in the budget, but if the central bank has to pay additional interest on reserves, it makes a smaller operating surplus and hence pays a smaller dividend to the government. Bank reserves may not be included in the total of public debt, but this is an accounting omission: these are a clear liability of the central bank and should be included in any comprehensive enumeration of public debt.

This is the left-hand pocket transferring something to the right-hand pocket misnomer.

The fact that the accounting records of the central bank records a lower “operating surplus” because they have provided more income to the non-government sector (interest on reserves) does not reduce its capacity to credit bank accounts (or load up vans with cash).

It wouldn’t even matter if the central bank recorded a perpetual accounting loss and had negative capital. The currency issuer cannot go broke.

Further, the fact that the treasury then records a lower surplus or higher deficit because the accounting flows from the left-pocket don’t appear in the right-pocket is also largely irrelevant for assessing the capacity of the government to net spend in the next period.

Remember the public debt at any point in time is just the accounting record of past fiscal deficits that have not yet been taxed away.

It is true that if we want to consolidate the liabilities of the government sector overall then bank reserves would be added to outstanding public debt.

But what would that tell us that was interesting? Not much at all.

Please read the following blog posts (among others) for more discussion of these themes:

1. The ECB cannot go broke – get over it (May 11, 2012).

2. The US Federal Reserve is on the brink of insolvency (not!) (November 18, 2010).

3. Better off studying the mating habits of frogs (September 14, 2011).

4. Central bank independence – another faux agenda (May 26, 2010).

5. The consolidated government – treasury and central bank (August 20, 2010).

Conclusion

So you see that:

1. The public debate is indeed shifting and Op Ed contributions from otherwise mainstream commentators are starting to invoke the frames that MMT has introduced to the public sphere.

You will not find analysis such as offered by Stephen Grenfell in any conventional money and banking textbook.

So that is a positive trend. Part of getting people to reach an MMT understanding is to get them to use MMT frames and language.

2. The characterisation of MMT in the public debate by those intent on criticism remains facile and error ridden.

The interesting point about this article is that in trying to put MMT down, Stephen Grenfell basically rehearses key MMT concepts without even knowing it.

3. Tying those operational realities in with Milton Friedman’s free lunch arguments, however, fails badly.

When Stephen Grenfell closes his Op Ed article with “But Milton Friedman got this part right: “there is no free lunch” – he is demonstrating a wilful misunderstanding of what MMT is about and what the ‘free lunch’ concept was about.

Paying interest on excess reserves is not constrained by scarcity! His analogy fails at the most elemental level.

That is enough for today!

(c) Copyright 2019 William Mitchell. All Rights Reserved.