« Help! The senior guy in my field acts like a total jerk sometimes. What should I do? | Main | Will legalization reduce youth marijuana use? »


Feed You can follow this conversation by subscribing to the comment feed for this post.


I already answered this question at the other thread, but I think you are seeking broader input, and understandably so. The Eurosystem is complicated.

That said, I think the answer to this is quite simple in concept, so I’ll try to paraphrase what I said before.

Yes, there is a structural similarity between a single central bank like the Bank of Canada and its set of orbiting Canadian commercial banks with that of the ECB and its set of orbiting national central banks.

But the two systems are very different for other reasons.

The Bank of Canada is the central bank for a set of capitalist commercial banks. That means that the pricing of imbalances needs to reflect that capitalist motivation in order for markets to be efficient. You know the rest.

The ECB is the super-central bank for a set of national central banks that have no similar independent direct capitalist motivation as institutions. In fact, it’s the opposite. The NCBs and their respective national governments each share in the net interest margin effect (or profit) of the entire Eurosystem of central banks in aggregate – according to their “capital key” or capital allocation shares. So there is no motivation for NCBs to clear imbalances with each for reasons of “market efficiency” – because the net result of the group all comes back to the same allocation to each national central bank of that aggregate result, regardless of the existence, distribution, and pricing of TARGET2 imbalances. Specifically, the pricing of inter-NCB imbalances has no net effect on aggregate or distributed profits because all of that pricing nets out within the group.

The system is designed to allow for such passive imbalances in the normal course. The expectation is that private capital flows working in an efficient market will minimize such imbalances on a cumulative basis. That obviously hasn’t happened. But that’s a flaw in the fundamental design of a system with a common currency but diffused fiscal policies. And the result has been a financial disaster with the ever lurking threat of national exits from the Eurozone. And such exits if they occur are where the resolution of these TARGET2 positions could really become a more serious issue than what was intended in the original design of the system.

I think you can see where this is going. I don't understand how it is supposed to work.

Why isn't the answer just that the liabilities underpinning Target2 balances require collateral and collateral haircuts (required by the national central bank of its domestic commercial bank)? Sure, these may ultimately prove insufficient in some cases and blow out transversality, but what else is new, right? That's still the basic constraint on the gross money stock rising without limit, right?

JKH and dlr: very good on-the-ball comments.

I think I'm going to let the comments flow for a bit, maybe letting the commenters argue among themselves, while I mull it over (and teach my class!).

The national central banks aren't really independent. They take orders from the ECB. At least, the Greek central bank did, at a time when it was painful to do so.

Note that in the event of a unilateral national central bank exit/default, the ECB would have the recourse of confiscating the NCB's equity in the ECB. So up to a point, the ECB is protected from that (remote) risk.

> The real world is a red/green world. It has both green money and red money. A positive balance in a chequing account is green money. A negative balance in a chequing account is red money. If I sell my car to Andy, who gives me a cheque for $1,000, the bank reduces my overdraft by $1,000 and increases Andy's overdraft by $1,000. The bank has transferred $1,000 of red money from me to Andy.

I disagree. The real world is not a red/green world. Your scenario violates the combination of two constraints from the monochromatic worlds:

> Only the issuer of green money is allowed to create green money.
> Only the issuer of red money is allowed to destroy red money. [edited to fix typo NR]

In the car-sale you describe, what has happened is that Andy's cheque is a transfer of strictly green money. If you did not have an overdraft, you'd have +1000$G in your account after the deposit. The fact that your overdraft is reduced instead comes about only because your commercial bank will create or destroy 1$G+1$R as required, up to your overdraft limit that is for now irrelevant.

My overdraft does not transfer to anyone else, because the medium of exchange is strictly green-denominated. Every instrument used to paper over the difference between a balance and an overdraft – cheques, debit, credit – involves somebody taking on the counterparty risk. Credit is a perfect example, in that if I purchase your car via credit card and then default on that debt, you are completely insulated; my card issuer needs to record a loss to reconcile its books.

The Red+Green world is still useful, however, because it tells us what happens when there are secondary issuers of currency. The point of a banking system is that my RBC chequing account can be spent like cash even if its balance is the result of a loan or overdraft. A credit crisis, where the secondary banking system decides that the existing set of overdraft limits is too high, acts to reduce the effective money supply just as if the central bank had tightened.

Majro: "> Only the issuer of red money is allowed to create red money."

Was that a typo? Did you mean to write "> Only the issuer of red money is allowed to **destroy** red money."? (Which is what I said)

(If it was a typo, I will edit your comment to fix it.)

As long as limits on red money holdings are set somewhere in the system won't this limit total Gross Money even at a higher level ?

For example: Suppose A and B have their own CB that sets a max red money balance of 100. C and D have a different CB with its own money that sets a similar limit. As long as the 2 CBs maintain their red-balance rule the maximum gross money has an upper limit of 200 per CB, and 400 in total. This can never be breached no matter what trades take place between AorB and CorD.

If the 2 CBs decide to form a currency union with each CB having accounts with a new SuperCB then isn't the amount of Gross money in the system still bounded by the red-money limits set at the CB level (as long as all trades are initiated by A.b.C or d and and not the CBs on their own behalf)? The CB for AandB could end up with a red money balance of 200 at the SuperCB and the CB for green CandD could end up with a green money balance of 200 , but even if the SuperCB has no limit the Gross Money supply can never go beyond 400.

1.If the Bank of Canada acted differently, and set the same rate of interest on both red and green money, so the spread is zero, then there would be no incentive for BMO and TD to trade in the overnight market.

That would indeed kill the overnight market. But it wouldn't have any effect per se on the amount of credit that commercial banks extended to their customers, would it? It would just become like your second point, just that the threshold is 0 for not lending. The CB would have to do the netting itself. Otoh banks' inter-connectedness would be eliminated which in turn would decrease systemic risk. Warren Mosler is a fan of eliminating the inter bank market for that reason, I believe.

What do you mean exactly by an 'implicit limit' on overdrafts?

Nick's dad was a trusted, faithful customer of the bank. The bank(er) may have grown rich over the years, partly due to the efforts of Nick's dad who repaid his green/red loan every year with green/red money.

How could a situation arise that a bank would have the ability to control Nick's dad's life with requirements to borrow and repay green/red money every year?

I turn to the concept of a national-gift-certificate. The owner of a nation/store can pay workers with gift-certificates. The gift-certificate becomes green money. The notation that a gift-certificate has been issued becomes red money.

Of course the owner of the nation would be better known as the National Government. Only the National Government has the ability to create legally accepted money/gift-certificates. The need for a medium of exchange that holds value over time is so great that Nick's dad can become dependent upon those who control issuance of money/gift-certificates.

Could this be the way it actually works?


“Eurozone national banks have chequing accounts at the European Central Bank.”

I saw your twitter exchange with Whelan, and I think you’re more right than he is.

If the Bundesbank buys French bonds from a seller who gets credit in a French bank deposit account, that shifts TARGET2 positive balances from the Bundesbank to the Bank of France, other things equal. So it’s analogous to BMO buying a bond from a seller who banks with Royal, with reserves moving over in the same way.

That said, in the same way that a BMO bond purchase from a seller who happens to bank with BMO has no net effect on the reserve account, a Bundesbank bond purchase from a seller who banks with a German bank has no net effect on the Bundesbank TARGET2 position. So the chequing account function is less visible in that type of example, and he may not be seeing the analogy because of that and because of the large scope of internal transactions within a particular NCB sub-system.

He refers to the Interdistrict Settlement Accounts for the regional Feds. That *is* a mechanism that resolves imbalances, but not for reasons of the type of market efficiency I was referring to in my first comment.

Whelan in my view was on the wrong side of the debates with Sinn several years ago.

... and so one question is whether the NCBs actively use that chequing facility with the objective of managing their TARGET2 positions in an active way

and the answer is no - due to the deliberate design and economics inherent in the transfer pricing of balances and the allocation of system wide central banking profits to the Eurosystem NCB participants and their respective fiscal authorities

@Nick Rowe:

> Was that a typo? Did you mean to write "> Only the issuer of red money is allowed to **destroy** red money."? (Which is what I said)

Yes, that was a typo. I'd appreciate the edit. The opposite is also true in the coloured worlds, where only the issuing bank can destroy green money and create red money, but since those decrease the holder's net worth in isolation there'd be no demand for it.

Another way to look at things with bank-issued money is who has to go to their bank to facilitate a trade.

In the green world, if nobody has money but a purchaser wants to buy something, they have to go to their bank to arrange a loan.

In the red world, if nobody has any money (0 balances) and a seller wants to sell something, they have to go to their bank.

In the red+green world, either option is possible.

In the real world, the buyer needs to go to their bank to facilitate a transaction if there's insufficient money at hand. This tells me we operate in a green world, but the easy access to credit means we have a lot of secondary issuers of money (Beta Banks).

Regarding TARGET2, I like the analogy with commercial banks and reserve accounts. With the positive/negative spread and interbank lending, there's a small but persistent incentive for banks to match their individual transaction flows. Since TARGET2 has no such mechanism, it means that the Bundesbank accumulates a diffuse asset and the Banca d'Italia accumulates an equivalently nebulous liability.

If there was a way to implement some sort of inter-central-bank exchange, then the nebulous liabilities would become concrete. I'm not sure if that would really help, though, as it would just seem to reimport the gold standard (with Serious People being concerned about flows of gold), only using the Euro as the new AU.

If we decide to call any negative balance red money, as Nick is wont to do, then the national negative T2 balances are red money. And of course Nick keeps pointing out in a red money system, there is the incentive to accumulate as large a red money balance as is possible before departing our glorious Earth, unless of course there is some means of sanction which dissuades from the accumulation of red money over the long term. This seems to be what's happening in the EZ clearing system. There is no incentive to reduce red money balances. These T2 balances are in effect the capital and current account balances that would normally accrue between nations in goods trading and financial transactions. They would normally be settled by formal debt arrangements or the exchange of foreign currency. There would also be pressure to alter the course of these imbalances. Deficit countries would eventually be compelled to deflate. In the Euro system, there does not appear to be the economic imperative to do this. The EC can only threaten to send in the Troika and cut off funds as happened in the case of Greece. In the case of Italy for instance, it's a case of too big to fail or too big to pressure and cause to fail.

T2 balance table also probably reflects the effects of the ECB's QE programme, with Germany being the main beneficiary.

"These T2 balances are in effect the capital and current account balances that would normally accrue between nations in goods trading and financial transactions."

Not quite.

They reflect only the extent to which the private sector fails to recycle such flows. There are some Eurozone countries where private sector recycling is reasonably efficient, with muted TARGET2 imbalance results.

"also probably reflects the effects of the ECB's QE programme, with Germany being the main beneficiary"

Not quite.

QE is a centralized strategy that is almost entirely decentralized in implementation via NCB execution and booking. There is no TARGET2 disruption when the Bundesbank purchases securities from German sellers. That largely immunizes against the kind of disruption to TARGET2 imbalances that might otherwise result with implementation and booking at a central point.

JKH: Here is a link to a Bloomburg article (dated 2016-09-21) that covers the increase in TARGET2 balances in more detail.

Thanks Roger.

The linked ECB speech there is quite revealing in detail:

“It follows that when a national central bank purchases securities from a non-domestic bank, the purchase is likely to give rise to a cross-border flow of central bank money. As around 80% of Eurosystem purchases by volume have been carried out through non-domestic counterparties, cross-border flows of central bank money are happening on a large scale as a direct result of the asset purchase programme… so when Banco de España purchases securities from a German counterparty, or from a UK-based counterparty that participates in TARGET2 via Germany, it can give rise to a flow of central bank money from Spain to Germany and affect TARGET balances [Figure 2]. ECB staff estimate that around 60% of purchases by volume under the asset purchase programme have been made from counterparties that participate in TARGET2 via Germany, while only around 5% have been made from counterparties that participate via Spain. It is in this context that Germany’s TARGET claim has increased by around EUR 200 since end-2014 while Spain’s TARGET liability has increased by around EUR 125 billion over the same time period.”

I hadn’t realized there was so much cross-border purchasing by individual NCBs implementing QE. Haven’t been following it recently.

My previous comment was not entirely wrong, but Henry’s was much more right with respect to the effect on Bundesbank TARGET2 balances. Sorry Henry.

... that apparently was a change in the pattern from direct NCB effects on TARGET2, as the Bloomberg article says:

"In the run up to mid-2012, peripheral central banks were lending to banks in their own countries and those banks were then using that money to repay debts to banks in the core."

Good comment thread. I'm still reading and mulling.

Anyone got any thoughts on dlr's comment above? If the red money is always matched with colateral (my father's bank manager won't let him run an overdraft bigger than the value of the farm) that sets a limit.


My comment at the other thread on the credit risk/collateral issue:


Is saying "the red money is always matched with collateral" more or less the same as saying "The bank that issues the red money needs to put some limit on each customer's holdings of red money" (as you do in the post)? Even if the red money is sometimes issued without explicit collateral banks will probably limit red money issuance to what it believes it can get back from the borrower if push comes to shove.

But if everyone in a country obtained red money to the limits of their collateral (or credit worthiness) it would still be scary high number.

Regarding dlr's comment

My first impression was that the limiting process he refers to relates to the growth in the gross money stock in the context of commercial banking money in a given NCB sphere.

I think that's not quite the same as the limiting process that relates to the growth in gross TARGET2 balances. It's certainly related - as an overall constraint on the origin of deposit migration from one NCB sphere to another. But its seems to me it's not quite the same thing as a constraint on the process of capital migration itself. But I'd have to think about it a bit more.

“There is a symmetry between the red and green worlds; one is the negative mirror-image of the other. But there is also one asymmetry: the red world has a fundamental problem. Each individual can increase his utility by buying more goods and selling less goods, thereby accumulating an infinitely large stock of red money. The bank that issues the red money needs to put some limit on each customer's holdings of red money, to ensure this does not happen. This is not a problem in the green world, because having zero stock of green money sets a natural limit that stops an individual buying, and the individual is fully aware of that limit.”

Of course, the same problem does exist in the green world for overdrafts and loans. Limits are still required.

(unless you prohibit overdrafts and loans - which I don't think you did because you refer to borrowers of green money)

The problem is essentially no different – just the form of the liability to which it applies.

let's say every italian depositor decided to move their money to german banks. and the italian banks replace these deposits with ECB borrowings. meanwhile the german banks receive all these new euro reserves and do who knows what. gross money has increased. what are the constraints to this increase in gross money outstanding?

first, the italian banks cannot do this indefinitely. they will only be able to borrow from the CB to the extent they can provide sufficient collateral and are deemed solvent. once every or more likely most (given haircuts and other obvious factors for anyone following SSM and italian banks) deposits are gone, they by definition will not be permitted to borrow (either directly or via the bank of italy) from the eurosystem. similarly, italian borrowers cannot take on overdrafts without limit for the same reason any other bank borrower can't take on overdrafts without limit. thus, gross money cannot increase "without limit."

second, the ecb's inflation target may act as even stricter constraint. if an increase in gross money outstanding develops due to country flight, the increase will generate tightening from the ECB even before hitting the italian bank/borrower capital limit constraint, if it results in euro-area inflation rising above target. if it doesn't, who cares that the gross money has increased? well, italy wouldn't be too happy that its banks were entirely wholesale CB borrowers and had no depositors, but that is just a standard story of what happens when a bank faces a deposit run, and not a story about gross red/green money increasing indefinitely.

of course, it is more likely that the increase in gross money does not spark inflation in the first place. this is because (1) it is a mistake to think that gross money can increase without limit and so there is no reason to expect a self perpetuating inflationary effect (2) the increase in euro reserves outstanding will be accompanied by a dramatic increase in volatile deposits. german banks would probably just hold more reserves, and also the shit hitting the fan which usually a flight to safety (3) it would be a mistake in general to focus on gross money balances outside of the CB reaction function which would presumably be unchanged.

i don't see how target2 is even interesting in this regard. when depositors started trotting away from WaMu and Indymac, those banks could replace deposits by borrowing from the Fed and paying those reserves to JP Morgan, i.e. the deposit recipient. just like T2, that can increase the gross money outstanding (and interbank lending was frozen so you can forget about netting effects spread or no spread) until WaMu and Indymac et al run out of borrowing base, which happens before infinity. so what.

Nick says:

“The ECB can hold the net money stock fixed, but the gross money stock can rise without limit”.

I take Nick’s intended meaning that there is no formal credit risk limit imposed in cases of TARGET2 deficit positions, or that there is no economic motivation to prevent an indefinite rise in such balances. I imagine Nick is quite aware that things generally stop before infinity is reached.

There is also a confusion in the comment discussion regarding gross money – a confusion that is not found in Nick’s post. Nick refers to gross money in the specific context of TARGET2 balances. In the broader context of Eurozone operations, that is most definitely not the same consideration as gross bank reserves. When the German banking system absorbs an inflow of money from the Italian banking system, it is a matter of standard function that the Bundesbank can take steps to drain those excess reserves from the German system. That requires no shift in ECB monetary policy. So that is not the same point as the one raised by Nick’s question about gross TARGET2 balances. Bundesbank reserve drains per se do not affect TARGET2.

I must say I’ve never understood how an elaborate explanation becomes proof that the elaborate is uninteresting.

From my perspective, this particular question is quite interesting because it revisits some very heated blogosphere discussions that took place several years ago on this very topic - or on questions that swirl around it. The big one is how does a TARGET2 deficit position get resolved if the country with the deficit exits the Eurozone. The EZ architecture, including transfer pricing of TARGET2 balances and allocation of aggregate Eurozone NCB profits, did not anticipate such an event.

JKH Nov. 26 at 01:02 PM :

If a country exits the Eurozone?

Am I correct that TARGET2 is recording the imbalances accumulated by the CB from each country? Am I still correct that the imbalances must equal zero within the TARGET2 balance sheet?

If this is correct, then the national CB's with positive balances should be lending to the national CB's with negative balances.

If my understanding is correct, a debtor country leaving would raise a repayment question for the creditor nations. This would be first a CB level problem, but private banks exist at the discretion of each CB, and individual bank depositors depend upon bank integrity. Repayment questions would reach all EU members and individuals.

Would the country leaving also be leaving the EU banking system? If so, I would conclude that it would be incumbent upon the leaving nation to establish a new banking system and attempt to reestablish trade patterns with other nations.

Those are good questions Roger.

I did a post on this 4 years ago.


Here's an excerpt (near the end) that touches on some of the risk issues. You may be interested in other sections as well.


Eurozone Risk Once Again
TARGET2 asset and liability positions reflect mismatches in the cross border flow of funds relative to normal patterns of private sector capital flows. These liquidity imbalances can indicate the reluctance of the domestic private sector (e.g. Germany) to assume foreign credit risk (interbank lending withdrawal), or risk averting behavior initiated by asset holders in other Eurozone jurisdictions who seek to move their assets (e.g. bank deposits) to a safer country (capital flight). This results in TARGET2 asset balances building up in “core” countries such as Germany and TARGET2 liability balances accumulating in the “periphery” countries such as Ireland, Greece, Spain, and Italy.

TARGET2 asset balances are associated with credit risk, but they do not incorporate credit risk. The related credit risk is found by tracking down the balance sheet assets of Eurozone national central banks and their commercial banking systems. For example, German commercial banks may withdraw from interbank lending to Spanish banks due to credit risk on the balance sheets of the latter group. Those banks turn to the Central Bank of Spain for replacement funding. This central bank replacement funding provided by the Bank of Spain absorbs some of the risk exposure that private sector depositors have fled. However, this risk is not for the account of the Spanish central bank alone. It is for the shared account of all 17 Eurozone national banks, according to the formula for such loss sharing by Eurozone members, based on the share of capital each NCB has invested in the ECB. And although TARGET2 asset-liability positions do not necessarily represent specific allocations of credit risk exposure, they do represent a type of intra-Eurozone liquidity imbalance that is attributable to private sector avoidance of risk in those areas whose central banks end up depending on TARGET2 funding.

A TARGET2 asset claim such the 727 billion euro position held by the Bundesbank (July 2012) has no contractual repayment characteristic. Its duration is indefinite. We simply don’t know how long it will be before this position is reversed. The existence of the position reflects a failure of proactive German capital outflows to recycle funding inflows on current or capital account. The German private sector is avoiding periphery risk (e.g. avoidance of interbank lending). And the periphery private sector in some cases is seeking shelter from their own domestic risk (e.g. cross border deposit outflows).

A future reversal of the German TARGET2 asset position would require a reversal of this general pattern of periphery risk avoidance. Given the current distribution of perceived risk across different Eurozone areas, it is unclear how or when such a reversal of capital flows would be forthcoming. A TARGET2 asset is extremely illiquid, due to its unknown duration and the challenge of extinguishing it through reversion to more normal financial flows.

We referred earlier to nonstandard transactions that the German central bank might itself undertake. For example, German TARGET2 asset balances might be “liquidated” in theory by direct exchange for actual reserves (or equivalent deposits) held by the Bundesbank with another Eurozone central bank. But actual reserves wouldn’t be much more useful than the TARGET2 balance (which can be interpreted as a loan of reserves). The German central bank as described earlier has the option in theory of purchasing foreign financial assets, which would create reserve credit for the account of another Eurozone NCB with a corresponding debit to the German TARGET2. Such a strategy would include taking on the credit risk specifically associated with such purchased assets. However, it is unclear how this would benefit Germany, since the result won’t necessarily change the credit risk exposure already existing on peripheral NCB balance sheets, for which the Bundesbank has shared accountability through the Eurosystem loss sharing formula.

It is not just TARGET2 asset holders who face credit risk exposure inherent in the ECB asset mix. All 17 national central banks face this risk, regardless of their individual TARGET2 positions. This is because any such losses are allocated according to the share that each of the 17 Eurozone members has in ECB capital. So the individual burden of credit risk exposure is the same, whether a particular TARGET2 asset is held by Germany or Luxembourg, for example. It is in this sense that there is a clear operational separation between the basic liquidity properties of a TARGET2 asset and the original credit risk exposure that may have been the risk catalyst for the creation of that asset.
But this intended separation of liquidity and credit risk according to Eurosystem design is not the end of the story. This is because in the worst credit risk scenarios, the intended Eurosystem function is itself at risk, meaning that the loss sharing formula is at risk. Germany’s outsized TARGET2 asset position indicates a deteriorating asymmetry of credit risk distribution across the Eurozone. While Germany is not exposed to a risk share that correlates exactly with the relative size of its TARGET2 asset, the size of its TARGET2 asset does indicate growth in the risk that it may be exposed to according to its committed share. So the fact that its TARGET2 asset is growing is not to be ignored.

And with growth in the depth and breadth of aggregate credit risk, there emerges the compounding complication that Germany’s formula share of loss absorption (currently about 27 per cent) may default into a de facto larger share, based on the risk that other Eurozone members may fall into circumstances in which they would not be able to meet their own loss share obligations. One must consider the scenario where larger events overtake intended system design, in which an NCB might end up defaulting on its shared obligation. This is conceivable when considering the possibility of Eurozone structural breakup, where a number of countries with outsized existing TARGET2 liability positions might leave the Eurosystem and fail to meet their liability obligations within it. Not only could the aggregate risk grow larger in such a scenario, but the number of countries capable of sharing in related losses could shrink as well. This suggests a potential negative convexity effect of accelerating aggregate losses combined with decreasing numbers of Eurozone members able to absorb them. Troubled countries may grow in number. And therefore the number of remaining countries capable of absorbing ever greater losses shrinks, and the share of those remaining countries in absorbing the losses grows. And this revised de facto loss distribution could be more concentrated than the original 17 member pro rata capital key formula.

Hans-Werner Sinn in his June 2011 paper described a TARGET2 “stealth bailout” of the periphery by the core. TARGET2 is not the immediate source of risk. But its imbalances reveal a flight from risk in peripheral asset portfolios. The Bank of Spain’s MRO asset positions for example transmit credit risk to the rest of the Eurozone, through the potential for losses and the subsequent activation of the loss sharing formula. The imbalance represented by Spain’s TARGET2 liability and Germany’s TARGET2 asset is symptomatic of German private sector avoidance of that risk. Critics of Sinn implicitly assume that the loss allocation and funding mechanism would function as intended under all circumstances, and that fiscal resolution of losses works according to formulae created under assumptions of continuing Eurosystem viability. This is a non-trivial assumption in the current environment.

In this connection, foreign exchange risk becomes a close relative of credit risk within the profile of comprehensive Eurozone risk. Some analogize the Eurozone as a fixed exchange rate system. The euro in this view is in effect a foreign currency for each of its 17 operational users, fixed by its original booking exchange rate to the predecessor currencies such as the Deutschemark. Structural foreign exchange risk may emerge with contingencies such as the departure of a member nation or broader Eurozone zone break up. Again, some analysts assume the ECB functions smoothly under such conditions, and that all operates according to plan regarding the allocation of risk and loss. But loss sharing formula itself may be at risk in the event of sufficiently disruptive losses.
Such a scenario probably underlies Sinn’s point about stealth bailout. Structural risk in the Euro system may include the possibility of exit and currency conversion for certain peripheral nations. This puts TARGET2 Euro claims of asset holders such as Germany at a new particular risk for a counterparty foreign exchange obligation. It creates an additional risk layer beyond that associated with a single currency. Indeed, some have suggested outcomes as radical as Germany itself exiting and converting to a new Deutschmark.

As already noted above, dismissing the relevance of central bank capital is generally short sighted. Central bank losses matter, despite the natural capacity of central banks to earn seigniorage income. And the popular argument that central bank solvency is not an issue because of built in recapitalization requirements also misses the point. It is mandatory that ECB capital must be replenished to required levels in case of losses. But this expected mechanism does not mitigate the importance of losses. The issue here is the substance of the fiscal connection between central banks and their sponsoring authorities. In the case of the Eurozone, the potential deficit impact is reflected across all 17 Eurozone government budgets in such a loss scenario, with the risk of more concentrated loss absorption than that in extreme outcomes, as described above.

The important thing about Target 2 is that it is about *refinancing* short term legacy debt. The (international) debt already exists - but private banks suddenly do not allow the debtor banks to roll over these existing debts anymore. The debtor banks however do pay (which the ECB wants to prevent the creditor banks from going bankrupt) and refinance this with a Target 2 overdraft. Mind that this is a net overdraft, the result of myriads of micro transactions! The ECB has to allow this because of its mandata (not the inflatin mandate, but the mandate to keep the payment system going. Again: the debts already existed (in the private sector).

An example is Italy. Italy has a current account surplus for a couple of years now. Despite this, its Target2 position is deteriorating - because legacy debt is not rolled over anymore by private banks. This is, however, implicitly financed by the Italian Central Bank. Which runs a Target 2 deficit because of this. As the Italian banks however paid the creditors, a matching surplus exists with the Centrla Baks of other countries. The point: there is no new net debt. Sinn is wrong (as is shown by the fact that Italy has a deteriorating Target 2 position despirte its current account surplus). And yes, Sinn is right that there will be a redenomination risk when the eurozone breaks up. As Goethe once stated: 'Intelligent people are sharpest when they are wrong'. The situation is indeed complicated by QE. But that's not what this discussion is about.

"Anyone got any thoughts on dlr's comment above? If the red money is always matched with colateral (my father's bank manager won't let him run an overdraft bigger than the value of the farm) that sets a limit."


Isn't the problem here that the overdraft is denominated in terms of green money (i.e. it has to be settled with green money) and then you want to call the overdraft red money.

The thing about red money is that it doesn't have to be settled at all. Isn't that correct? (That's why you keep saying that it's in everybodies interest to die with a mountain of red money.)

I am going to try to state it another way.

The movement between commercial banks and/or central banks is just that, movement. The banks are not trying to spend more than they earn. It is for their customers.

Entities spending more than their checking account contains are trying to spend more than they earn.

"Each individual can increase his utility by buying more goods and selling less goods, thereby accumulating an infinitely large stock of red money."

Henry, this could also be partially true?

Each individual can increase his utility by buying more goods and selling less goods, thereby accumulating an infinitely large stock of green bonds denominated in green money.

"My previous comment was not entirely wrong, but Henry’s was much more right with respect to the effect on Bundesbank TARGET2 balances. Sorry Henry."


Whatever I got wrong is because I don't really know what I am talking about. You actually know what you're talking about.

Whatever I got right is because I fluked it.

My first comment about T2 balances reflecting intra Eurosystem balance of payments is not entirely correct because:

1. the location of the payee/payer within Europe determines how each NCB's T2 balance is affected

2. some cross border (but within Euro system) payments made with commercial bank money ans some with central bank money (as you pointed out).

3.euro transactions involving non-euro area banks can be effected by accounts held by banks in euro area countries.

Thinking of the national central banks as commercial banks or even in any way distinct from the ECB just leads to a complete mess and is substantively quite wrong. Koning is right on target that the proper analogy is to Federal Reserve regional banks. As such the thing limiting money creation in the Eurozone is what's limiting (or not limiting in some cases) money creation in the U.S. or Canada- macro-prudential regulation and capital requirements. The difference is that unlike the commitment to preserve the stability of the payments system central banks have in the rest of the world, the ECB doesn't believe it has such a mandate. http://www.nakedcapitalism.com/2015/07/mario-draghi-the-ecb-has-no-mandate-to-ensure-checks-clear-or-credit-cards-work.html

This leads to the biggest messes outside of the denial of fiscal space to individual governments

Now if you're asking how is such a system supposed to preserve macroeconomic stability and growth- the answer is it isn't.

JKH: Very good post on TARGET2 four years ago! For what it's worth (myself being a non-expert), I agree with you.

Your wording here bothered me slightly: "...that is attributable to private sector avoidance of risk in those areas whose central banks end up depending on TARGET2 funding"

The way I understand this, I wouldn't say the (national) central banks (NCBs) depend on TARGET2 funding. Not to say it's wrong to say so, but I feel this needs some clarification. (Which I think you provide in your broader discussion, but I just want to make sure we agree on this.)

Isn't the NCB mainly a passive player in all this? If a Spanish person with a deposit in a local bank wants to convert it to a deposit in a German bank, then Spain's TARGET2 deficit (assuming it has one) will automatically grow when the transaction is processed (real-time gross settlement?). There's nothing Bank of Spain can do about it? That's why I wouldn't say Bank of Spain depends on TARGET2 funding.

Because of this kind of "deposit flight" and "roll-over flight" of private cross-border lending, the local commercial banks end up depending on NCB funding, and the NCB's TARGET2 deficit (or negative balance) grows -- I can see that this is funding for the NCB, but it's "build in the cake". Perhaps we could say that the whole national banking system depends on TARGET2 funding?

Was this what you had in your mind when you wrote the sentence?

I think they do depend on it Antti - in full institutional context.

When capital flees Spain to Germany, there is a reserve drain from the balance sheet of the Spanish NCB.

So by construction/deconstruction, it is short funding (in the form of reserves) for its assets.

There is a hole in its balance sheet that needs to be plugged.

Think of what happens if that hole isn’t plugged.

That means that the Bundesbank will have accepted new incoming liabilities (a reserve inflow) without being compensated for that with offsetting assets. And the Spanish NCB would have lost outgoing liabilities without having to replace them.

If that happened, the Spanish NCB would actually write up its equity and the German ECB would write down its equity. That sort of thing starts to matter when you think about things like interest margins and fiscal effects from CB profits, aspects which matter more and more if rates go back up again.

I.e. banks don’t like to take on new liabilities without being compensated with new assets.

So the ECB provides an offsetting asset to the Bundesbank – the TARGET2 asset.

And the ECB “forces” a new liability into the Spanish central bank – the TARGET2 liability.

And all of this happens through logical operational clearing of payments - which is the ground level effect of how what needs to happen does happen.

So the Spanish central bank depends on TARGET2 funding in the sense of making its balance sheet coherent in that way. That said – this is a dependency that’s forced on it – but for good reason and in somewhat the same way that a central bank LLR facility is forced on a commercial bank short of reserves (where there is a net liability in the reserve account – like red money (hello Nick)).

The cynic might say this is all bookkeeping.

The realist might say that the coherence of the monetary system depends on bookkeeping.

Having said all that, it’s only half the story in the Spain/Germany example.

The TARGET2 deficit funding for Spain plugs the hole from an equity coherence perspective.

But it doesn’t plug the hole from a reserve perspective.

The Spanish central bank must still take steps to ensure that the reserve drain due to the capital outflow is replaced in such a way to ensure continued functioning of its domestic banks. This can be done through some variation on OMO.

If the reserve loss is fully replaced, the balance sheet actually expands by the size of the TARGET2 deficit funding injection.

“There's nothing Bank of Spain can do about it? That's why I wouldn't say Bank of Spain depends on TARGET2 funding.”

Have a look at the first comment I made under this post:


One thought:

Suppose the Eurozone was the whole world (it isn't, of course). Then Target2 balances would (I think) be like the cumulative balance on the Official Account of the balance of payments (the sum of current surplus plus capital account surplus). Target2 balances would be like the foreign exchange reserve position of the national banks.

But with (say) gold reserves in the gold exchange standard, gold reserves can't go negative. But Target2 balances can go negative.

I think that's right.

Antti (& JKH)…

Depending on how broadly we define "funding", we actually have some examples where the ECB effectively forced the NCBs (and governments) to assume explicit responsibility over liquidity assistance to domestic banks.

The Irish central bank provided emergency liquidity assistance (ELA) to liquidity constrained domestic banks once it became clear they didn't have enough collateral by which they might have simply borrowed funds from the NCB. Trichet (the ECB chief) threatened to cut off all such funding unless the Irish government applied for a bailout; effectively involving the balance sheet of the country's public sector in the operation. The ECB also made Greek government bonds ineligible for OMOs and thus forced the Greek NCB and government into a similar kind of ELA-liability absorption.

The passivity of NCB-ECB funding, in regards to cross-border TARGET2 transfers/balances, is evident in cases where domestic banks have enough reserves or rather enough eligible collateral to fulfil cross-border transfers. But once there's a crisis, the whole situation becomes murkier, especially if ECB-approved eligible collateral becomes scarce. That's where "funding" turns out to also become contingent on how the ECBs assesses national funding schemes.


Wouldn't that cumulative deficit translate to a cumulative outflow of gold for international settlement on a gold standard?

And so couldn't the (residual) stock of gold reserves still be positive - notwithstanding a cumulative outflow of gold?

So I think the cumulative change that corresponds to the existing TARGET2 position can go negative under either fiat or gold. Under an international gold standard, that TARGET2 measure would not appear as a CB balance sheet liability the way it does now. It would only be the quantitative measure of the cumulative outflow of gold from gold reserves. And gold reserves would be less by that amount. But they might still be positive.

I could well be wrong though.

Now I'm wondering if what I just suggested partly depends on whether or not there is domestic gold production.

JKH said: "So the Spanish central bank depends on TARGET2 funding in the sense of making its balance sheet coherent in that way. That said – this is a dependency that’s forced on it"

Yes. The forcing part is what I meant with my mysterious "build in the cake" metaphor (where did I get that?! I even misspelled 'built'). I don't think we disagree on anything important, really.

One small thing bothers me slightly, though... You said: "There is a hole in its balance sheet that needs to be plugged." That makes it sound like it needs to be actively plugged, although as you said, it is plugged automatically. I know this sounds like nitpicking, and it is, but we should be careful when the 'need' we talk about refers to logic, not to something someone needs to do. (The reason why I even bother to mention this is that this is a real problem in discussions about sectoral balances, where you hear people say things like "households need to get into debt, because the public sector and businesses run surpluses" -- wrongly suggesting a clear, one-way causality, the direction of which iss known.)

Johan: Good point about ELA. I was talking about non-ELA situation, where you seem to agree that the NCB's role is more or less passive.


Good point

In this case I mean need in the sense of system design with respect to sensible balance sheet outcomes from the clearing process

Johan raises some good points. Although I’m not sure that they relate directly to the “passive” mechanism whereby TARGET2 balance outcomes are determined. As far as I know, there is no way of interfering with actual outcome for reserve balances that are cleared from one NCB jurisdiction to another – regardless of the precipitating events that cause the capital flows. My understanding is that crisis-specific specialized arrangements such as ELA are ring fenced in the sense of how their fiscal consequences are allocated – i.e. the fiscal risk is for the account of the country in question as opposed to being pooled with the rest of the EZ central bank results and distributed according to normal capital key allocations. But I’m a bit hazy on this aspect.

Merijan Knibbe writes "............Target 2 overdraft. Mind that this is a net overdraft, the result of myriads of micro transactions."

I think Merijan is making an important observation.

I don't know exactly how "myriads of micro transactions" can result in a national debt to a second nation but it seems to occur. The TARGET2 overdraft indicates that this is occurring but does not identify the mechanism.

This was happening in Greece. In Greece, it seems to have happened because the Greek government borrowed (and borrowed) so that it could pay generously. With generous pay, Greek citizens could buy German goods. The Greek government became a debt-enabler.

Is the same thing happening in Italy today? I don't know.

JKH: "And so couldn't the (residual) stock of gold reserves still be positive - notwithstanding a cumulative outflow of gold?"

Yes. It all depends how big the stock of gold reserves is when we start counting the cumulative losses.

The analogy is inexact, because with TARGET2 everyone starts from zero, and you can't mine TARGET2 balances like you can mine gold reserves, as well as gold reserves being unable to go negative.


"In this case I mean need in the sense of system design with respect to sensible balance sheet outcomes from the clearing process"

This was quite clear, and that's why I called my commenting on it nitpicking :-)

I very much like the way you focus on "sensible balance sheet outcomes".

My interpretation of Eurosystem TARGET2 balances is:

It is the accommodating item in the balance of payments as opposed to autonomous items. The latter happen independent of other items (to first approximation). Accommodating items are the opposite.

To be more accurate ... accumulated accommodating flows

This way of studying it ... accommodating item vs autonomous item ... is more accurate.

Gold sales for example is an accommodating item in the gold standard. But the stock of gold can't be negative. If a central bank loses the stock of gold, it can anyway borrow in foreign currency and make a sale of the proceeds. So that would be accommodating flows.

"as well as gold reserves being unable to go negative"

bearing in mind that cumulative gold flows can go negative, which is part of the comparison

(a comparison using a selected common starting point)

This way of defining is due to James Meade:


Page 83 onward.

accommodation = residual = net clearing result

in context of private versus "official"

dlr is right.

Here is an interesting paper which describes imbalances between regional Feds:


Isn't the problem here that the overdraft is denominated in terms of green money (i.e. it has to be settled with green money) and then you want to call the overdraft red money.

The thing about red money is that it doesn't have to be settled at all. Isn't that correct? (That's why you keep saying that it's in everybodies interest to die with a mountain of red money.)"


I would appreciate a comment on my earlier point.

Mr Sinn is at it again:


Sinn writes: "Frequently, if the sellers of Spanish government bonds are outside the eurozone, it will ask the ECB to credit the payment orders."

This caught my eye, and I thought this blog might be the right place to ask:

Is Sinn right on this one?

I'm thinking that where the sellers are domiciled isn't decisive here. What matters is through which banks they transact, and at least one of these is often in Germany -- or so I have understood. I'm not a fan of stories like the one Sinn comes up with (after the above quote), where the sellers, having sold the bonds, "transfer money" to Germany or the Netherlands to buy companies or company shares.

David Hume's "price-specie flow mechanism" was a theory that explained why there was an equilibrating mechanism that prevented all the money (gold & silver, "specie") ending up in one part of the world. But David Hume's theory applied to a green world.

I'm trying to figure out of there's anything equivalent to Hume's Price-specie flow equilibrating mechanism in a red/green world, like Target2.

"I'm trying to figure out of there's anything equivalent to Hume's Price-specie flow equilibrating mechanism in a red/green world, like Target2."

Nick, really interesting question.

Local Exchange Tradings Systems (LETS) are exactly like Target2 red/green money.

My understanding is that when LETS initially appeared in the 1980s, they allowed unlimited negative balances. To get around the problem of "all the money ending up in one part of the community," they eventually began to implement credit and debit caps. Also, when a certain period of time had passed, LETS members (both debtors and creditors) were often required to balance their accounts back to 0.

i.e. https://wiki.p2pfoundation.net/LETS#LETS_systems.27_major_weakness:_Debt_policing

Interestingly, Keynes' bancor plan was a red/green system. Here is Keynes:

"Measures would be necessary to prevent the piling up of credit and debit balances without limit, and the system would have failed in the long run if it did not possess sufficient capacity for self-equilibrium to prevent this. (Proposals for an International Currency (or clearing) Union, February 11, 1942)"

He devised a number of mechanism to prevent this, which I went into here:


So I don't think there is an automatic equilibrating mechanism in a red/green world like Target2. It needs to have that design feature implemented at the outset, like a LETS with credit limits or Keynes' bancor.

To some extent, T2 balances reflect national balance of payments. So a country with a T2 liability is experiencing, probably, a net outflow of funds owing to a current account deficit. In a fixed exchange rate system, diminishing foreign reserves pressure would eventually force the deficit country to deflate and staunch the net flow of funds outward. This was the equilibrating system under fixed exchange rates. Within the EZ (and it is effectively a fixed exchange rate system, one Irish euro equals one Spanish euro, for instance), this dos not appear to happen. Funds draining from a nation's system can be replaced by the NCB of that country. So that country's money supply does not fall even though its T2 balances are becoming less positive or more negative. That is, there are no deflationary forces in the deficit country to reduce the current account deficit. And as the ECB does not compel settlement of T2 balances, they go on increasing ad infinitum.

JP: "Local Exchange Tradings Systems (LETS) are exactly like Target2 red/green money."

I wondered if they might be.

Really good comment.

And your excellent post anticipates a lot of what I'm talking about here.


Willem Buiter did an analysis of the issue once which in my view was too alarmist as the situation is actually under control of the governing council of the ECB:

JP: Great post on ICU/EMU, thanks! Your last paragraph summed up my thoughts as well, although in a much clearer way than I could have ever put it. Reading your post, it crossed my mind that no wonder many gave up on thinking how to govern the (international) financial system, and assumed instead that the "markets" will govern it on their own ;-) As you suggested (I think), the governance might work on paper but in practice the plan cannot be put to work.

From one of your quotes: "Each member-State, on joining the system, shall agree to pay to the Clearing Union any payments due from it to a country in default towards the discharge of the latter's debit balance and to accept this arrangement in the event of falling into default itself."

This reminds me of the bilateral clearing arrangements Germany had in place during the 1930s. The Magician himself, Hjalmar Schacht, wrote about this in his book "The Magic of Money" (freely available as PDF). He said that Germany couldn't accept money as payment for its exports, because that money would have gone to debt repayments as per agreement with the creditor nations (who also had at least partial control over the payments system) -- and this would have meant that Germany's imports would have been restricted to bare essentials (as defined by creditors).

Instead, by exporting Germany earned credits in the bilateral clearing institution's books (these were many, one for each nation trading with Germany), while the importer incurred debits. These balances were then cleared by German imports from the same nation.

Going to back to Keynes quote above... His 'payments' probably meant "payments in money", and the "Schacht way" of getting paid in kind -- Schacht called this barter -- would have still been open for debtor nations? One cannot force sovereign states to stick to contracts if they themselves see that doing that would cause undue harm for the nation and its populace.

(From bilateral clearing we, of course, get to multilateral clearing, and that kind of system should probably still be called a barter -- not monetary -- system. Which makes one wonder if our current banking system is, in fact, more like a barter system than a monetary system.)

"....we, of course, get to multilateral clearing, and that kind of system should probably still be called a barter -- not monetary -- system."


If you read Schumacher's (who developed the Bancor concept along with Keynes) paper "Multilateral Clearing", you will be left in no doubt that the clearing system is monetary.

Henry: I think I will continue to see the system as a barter system, but that's because I think that the "monetary systems" we have are all essentially barter systems. Jevons described the English banking system in the 1800s as being on its way towards a "complicated and perfected system of barter", and it's that kind of definition of 'barter' I have in mind.

But thanks for the tip, I'll read the paper! (I recall I've read something similar, but perhaps not that one...)

Excellent post by JP Koning.

Although it touches on issues beyond Nick’s original question about the difference between the economics of ECB clearing of national central bank positions and that of conventional central bank clearing of commercial bank positions.

While gross TARGET2 balances are large, net balances are close to zero. (I think a few quirks about EZ non-members transacting with the system account for this.)

So in comparing the ECB system to a regular CB system, one might analogize it to the case of the pre-2008 Fed, in which net reserve settlement balances were kept fairly minimal. So an NCB holding surplus or deficit TARGET2 balances at the ECB analogizes to a US commercial bank hold surplus or deficit reserve settlement balances at the Fed. However, there is a big difference in the capacity of an NCB to respond to its situation compared to a bank like JP Morgan operating in that pre-2008 environment.

An NCB surplus or deficit TARGET2 balance reflects a net balance of payments that has not been cleared in the other direction by the domestic private sector. Although, if permissible, an NCB could in theory change the TARGET2 position that is otherwise delivered to it by engaging in international transactions in Euros on its own account, there is no way it can *directly* change the official balance of payments position that creates that initial TARGET2 position. For example, if permitted, the Bundesbank could buy foreign assets denominated in Euros in an attempt to clear its TARGET2 surplus position. But that does not change the fact that the same net capital inflow to Germany is still being recycled by an official entity.

Furthermore, as I noted, the collective bottom line of the NCBs in total is for the accounts of the participating members according to a capital key formula. So any proactive initiative in NCB balance sheet management is for the account of all – not just the initiating NCB. There is something circular in this. If capital is fleeing from Spain to Germany, then what is actually achieved by the Bundesbank buying Spanish bonds that isn’t already achieved by TARGET2 clearing? Perhaps more return along with more risk? But the result is for the account of all NCBs and their national governments. Why bother? Moreover, that sort of NCB specific capital flow intervention can add significantly to the financial risks for the EZ group that may be associated with the contingency of Eurozone membership exit by the weak member that is the source of such capital flows. (This corresponds I think to the convertibility risk that JP references). Again, why bother to clear the initial TARGET2 balance position, if this is doing nothing directly to restore the healthy functioning of private sector finance?

By contrast, the position of JP Morgan with respect to a surplus or deficit at the pre-2008 Fed is much different. JP Morgan has no such profit sharing arrangement with other commercial banks that clear through the Fed. It does what it does for its own account only. To the degree that has that profit motivation, then net balances either way at the Fed reflect an inefficient use of its balance sheet. So it will immediately try to correct that position. Importantly, like all big commercial banks, there is a myriad of asset-liability adjustments that a commercial bank can undertake in order to correct its reserve imbalance and improve its return on assets and equity. (There is a tendency by economists not to understand this aspect of commercial bank reserve management. The typical analysis always seems to be that individual banks undertake corrective action through the interbank market. That is only one way in which this is done. In fact there are a range of responses that include asset and liability pricing adjustments across many categories, with corresponding elasticities of flows and balances and corresponding effects on reserve positions.)

So there is an economic incentive for individual commercial banks to minimize their surplus and deficit reserve positions, such that the “balance of payments” imbalance between banks is minimized. In the pre-2008 era, individual banks simply got rid of their excess reserves through a variety of means. And while the bid-offer spread in the interbank market helped, there are many other avenues of balance sheet adjustment, including liability management of deposit pricing with non-bank wholesale counterparties or managing liquid asset positions with such short term assets as treasury bills and commercial paper or liquid bonds. Even in the case of a failing bank whose deposits start to flee to other banks, the US financial system is sufficiently deep in order for the receiving banks to expand their assets and/or reduce other liabilities such that the Fed can still have tight control over the excess reserve setting while providing LLR to a near failing bank and restoring balance to the reserve positions of the healthy functioning banks.

Again, the ECB / Euro System arrangement is one of shared fiscal impact from multiple national central bank operations. So even if national central banks took steps to correct imbalances generated otherwise by the private sector, the balance sheet effect and therefore the fiscal effect of the resulting NCB portfolio earnings would be shared among the member countries according to their capital share in the ECB. I think that means there is little incentive for individual NCBs to undertake corrective action when it comes to international imbalances.

And I think this is all somewhat separate from Keynes’ idea of motivating current account and net capital flow adjustment with a penalty on imbalances. In the case of the Eurozone, that sort of problem is all deeply behind the scenes as the source of the impact on NCB TARGET2 positions. It really can’t be resolved in a direct way by the NCBs or in a way that is comparable to, say, commercial banks operating in the pre-2008 Fed system.

@ JP, I'll join in the congratulatory party, too. Now I understand Nick's post better.
I particularly liked your conclusion that a better system could not have been realised for political reasons. I think the EU should have built on the ECU instead of replacing it with the Euro. Then again, doing economics in subjunctive is cheating.

JP Koning: "So I don't think there is an automatic equilibrating mechanism in a red/green world like Target2. "

What eventually forces a foreign exchange current account deficit country in a fixed exchange rate system to redress the balance is the fact that it is losing foreign exchange reserves. This cannot go on forever as foreign reserve holdings are finite. The deficit country cannot create foreign exchange reserves/money - it can only create domestic money.

The problem with the EZ is that NCBs within the eurosystem can create "foreign" money as well as domestic money - because they happen to be of the same denomination. So if the country is in deficit with other euro countries it does not suffer the reserve constraint in a normal foreign exchange/currency system. It can happily continue to create money to cover the losses to its internal system - there is therefore no deflationary pressure serving to rebalance the deficit country's eurosystem payments.

Its not just then about setting credit/debit limits within the T2 system. It's about the ability of the eurosystem NCBs being able to create euro denominated reserves. If this was not possible, then NCBs would be disciplined by a reserve constraint.

JKH: "While gross TARGET2 balances are large, net balances are close to zero. (I think a few quirks about EZ non-members transacting with the system account for this.)"

Small point. And I'm not 100% sure about this, but I think that's not it. It might be because because the ECB itself has a Target2 account, and has used it to buy bonds (OMO/QE). (I haven't checked this out or thought it through, because it raises a host of other issues.)

Antti: "I'm thinking that where the sellers are domiciled isn't decisive here. What matters is through which banks they transact,..."

That sounds right to me.

I have forgotten who this comment is for:

My house, and the positive balance in my chequing account, are both assets. But only the latter is (green) money, because I can use it to buy things directly. (I can only buy things with my house indirectly, by selling it, or borrowing against it.)

My mortgage, and the negative balance in my chequing account, are both liabilities. But only the latter is (red) money, because I can use it to sell things directly. (I can only sell things with my mortgage indirectly, by buying it back.)


We may both be right.

This from an earlier Whelan paper:

“TARGET2 balances among participating central banks sum to zero. The TARGET2 balances of the NCBs and ECB summed to zero in 2006 but were non-zero in later years because countries outside EMU joined TARGET2, using it to process euro-denominated payments.”

And this from somebody else, more recent, concerning QE:

“The QE program has been decided on by the ECB, but is run by the Eurosystem. While most of the asset purchases are executed by the NCBs, not all of them are. The ECB itself also buys various assets (everything from sovereign bonds to covered bonds, ABS, corporate bonds and agency bonds). The ECB’s purchases are distributed across the Eurozone. So what happens when the ECB e.g. purchases €100 m. in agency bonds in Spain? … So the ECB’s balance sheet will now show the Agency bond as an additional asset and the negative TARGET balance as an offsetting liability.”

From JKH's last quote: "So what happens when the ECB e.g. purchases €100 m. in agency bonds in Spain? … So the ECB’s balance sheet will now show the Agency bond as an additional asset and the negative TARGET balance as an offsetting liability."

Just to make sure I've understood this (thinking out loud):

We can think of the ECB's own TARGET2 liability as (NCB) reserves. When ECB buys a bond, some private bank's reserve account is credited in the NCB ledger. At the same time, the NCB debits its TARGET2 account (a debit entry increases assets), which corresponds with ECB crediting its TARGET2 account (a credit entry always increases liabilities).

Have I understood correctly that some entities have reserve accounts directly in the ECB ledger? So that if the ECB bought the bond from/through this kind of entity, it wouldn't affect TARGET2?

I said: "We can think of the ECB's own TARGET2 liability as (NCB) reserves."

I should have added "kind of reserves". ECB's TARGET2 liability is obviously not an asset to any specific NCB (which makes it "red money", Nick?). I had in mind the connection to reserves in the NCB ledger. (Perhaps I should have kept my thoughts to myself...)

" But only the latter is (red) money, because I can use it to sell things directly."


This is cannot be correct. Your mortgage and your negative cheque account are exactly the same. You would have has much chance/facility using your cheque account to sell a motor car as you would using your mortgage.

Henry said: "Your mortgage and your negative cheque account are exactly the same."

I agree. I would say they are both a case of loans create (demand) deposits.

What happens if the reserve requirement is 100% everywhere in the EU?


“We can think of the ECB's own TARGET2 liability as (NCB) reserves”

This is a general comment I wrote 4 years ago in my post:


TARGET2 Balances as “Supra-Reserves”

An alternative view interprets TARGET2 balances held by Eurozone central banks with the ECB as analogous to reserve balances held by commercial banks with their central bank. In this framing, TARGET2 asset balances held by NCBs are net positive settlement balances of a “higher order” form of reserve. TARGET2 liability positions become deficit (borrowed) positions in these higher order reserves. This “higher order” form of reserve might be identified as a special type of reserve held by national central banks with the ECB, or a “supra-reserve”. This concept abstracts from the regular “lower order” case in which commercial banks hold reserves with a single central bank.

Within a single central bank system, a commercial bank payment includes the transfer of a reserve balance from payer to payee commercial bank. If the payer bank starts from a zero reserve position, it incurs a liability to the central bank for reserves owed (overdraft or borrowed reserves). The payee commercial bank will increase its reserve balance. Because the debiting and crediting central bank is the same bank, there is no net change in the central bank’s own reserve liability balance.

Within a multiple central bank system, a cross-border commercial bank payment will include a similar redistribution of reserve balances. The payer commercial bank’s reserve account will be debited by its central bank and the payee’s account credited by its central bank. But debiting and crediting central banks are different in this case. One will lose reserve liability balances; the other will gain them. As described above in the case of the Eurosystem, the payer central bank will incur a TARGET2 liability to offset its reserve liability reduction and the payee central bank will gain a TARGET2 asset to offset its reserve liability increase.

The relationship between the ECB and multiple NCBs in the clearing process for TARGET2 balances is similar to that between a single central bank and its commercial banks in the clearing process for regular reserve balances. A payment between two commercial banks in a single system results in a reserve decrease and a reserve increase. A payment between two central banks in a multiple CB system results in a TARGET2 liability and a TARGET2 asset. Thus, TARGET2 balances in a multiple CB system are functionally analogous to regular reserves in a single CB system. They both provide a clearing and settlement function for their own mode of inter-institutional payment.

The ECB acts as the clearinghouse for initial bilateral entries of TARGET2 balances, which sum to zero for the 17 central banks of the Eurozone. This zero sum characteristic is analogous to a regular commercial bank reserve system with a zero reserve requirement (e.g. Canada). Some banks hold positive balances while others borrow balances. The net reserves of those who have borrowed are zero, individually and collectively, while those with positive balances collectively hold the system excess reserve position that results from the uneven system reserve distribution.


As noted in my previous comment here, there are exceptions that result in a non-zero net balance for TARGET2 on the books of the ECB – including certain non EZ members holding balances and some monetary policy transactions carried out by the ECB for its own balance sheet rather than by the NCBs for theirs. In any interpretation, I would tend to think of the first type as “foreign” CB Euro operating accounts rather than EZ NCB quasi-reserve accounts. And while I was aware of the second type (I think mostly associated with special late crisis ECB intervention actions such as QE, as mentioned by Nick), I just haven’t followed those closely enough recently to know the full detail. The published detail of the ECB balance sheet (when you can find an unconsolidated balance sheet presentation) is pretty messy in its full classification.

sorry, spelling - Antti

Henry & TMF: I mostly agree with you, but at the same time I see more sense in what Nick is saying than you guys do. Please allow me to promote my blog post on overdrafts also in this thread. It explains how I see overdrafts and "traditional bank loans", and how I think Nick sees them. Let me know if you disagree on something I say?

JKH: Thanks again, JKH! I'll look into your answer when I have more time later today.

Suppose I have a term deposit and a positive balance in my chequing account. Both are an asset to me and a liability of the bank. But only the second is (green) money.

Suppose I have a bank loan and an overdraft in my chequing account. Both are a liability to me and an asset of the bank. But only the second is (red) money.

If I buy/sell a car by cheque, the second disappears directly. It takes an additional step for the first to disappear.

There's a symmetry here.


I think you’ve described a dual green/red money world where there is symmetry in the relative liquidity of green money positive balances and overdrafts (i.e. “red money” balances), compared to term deposits and loans.

I agree.

But there’s a different asymmetry in your example in the sense that the same payment instrument is being used in both cases (at least I infer that in this example).

It is positive green money balances that are being transferred by that payment.

In that dual green/red world, green money payment instruments can be used to adjust either green or red balances, whether they be source or destination of the payment, whether up or down in balance, whether positive or negative to start or finish.

The dual green/red world is essentially continuous at balance 0 with respect to the facility to make payments using what is essentially a green money instrument.

The asymmetry in your example is that if I sell something, it is still an incoming green money cheque that reduces my red money balance, not an outgoing red money cheque.

Maybe I misunderstood or have forgotten, but I think your original idea was that red money existed as a form of payment as well, where the direction of the payment flow in a purely red money world was in the direction of seller to buyer rather than the normal green buyer to seller direction.

In this case, having chosen an example in the dual world, you’ve made a choice about the payment instrument, which is biased to green. That’s the asymmetry.

JKH: think of the cheque that the buyer of my car gives me as an instruction to the bank to transfer red money from my account to his account.

you can think of it that way Nick - as an equivalent outcome

but in fact the cheque instructs a green money transfer in the other direction

(pay to the order of ... etc.)

again, I'm referring to the aspect of the choice of payment type in a dual world

as opposed to what that would look like in a red only world

in a red only world, the buyer could still write the cheque

but the instruction would be to pay a red balance from (rather than a green balance to) the account of the cheque recipient

JKH: In a red/green world, I think we can think about it either way.

Nick & JKH:

Nick said: "If I buy/sell a car by cheque, the second disappears directly. It takes an additional step for the first to disappear."

Did you have a look at my post? There I say:

"For instance, [in a world with only overdraft-type of "bank loans"] we would repay debt when we sell goods (item #3 in my document), not when the overdraft/credit limit is reduced (item #4). The latter is how we see repayment when we think of a traditional loan. If the accounting treatments are identical in their outcome, then why this discrepancy?"

To me, the "additional step" you talk about has no substance behind it, in the sense that we are talking about a choice regarding accounting treatment which should have no real-life (that is, outside the "accounting realm") relevance. That additional step should not be regarded as something having to do with the repayment of debt; it is actually equivalent to closing an overdraft line of credit. This might come as a surprise to many; at least it came to me when I realized it (don't remember exactly when -- perhaps 1-2 years ago).

With all due respect -- and that's a lot of respect in case of you two -- I think your discussion about whether it is green or red money that is transferred is, in the end, quite meaningless. All we have is a debit entry on one account and a credit entry on another account. Which comes first, debit or credit, doesn't matter (in your language: the "money" can arrive at the other account before it leaves the first). You're free to view it as a transfer of green or red money if that helps you in making sense of the system, but because that "transfer" is just a way to interpret those two entries, we can as well conclude that there is no transfer at all.


(1) the seller's liability (recorded in the bank ledger) is decreased, and

(2) the buyer's
(2a) liability is increased, or
(2b) credits/assets are decreased.

Of course, we need to explain the meaning of those liabilities/debts and assets/credits. That's what I'm doing with my blog posts on "impure gift economics". These are records of gifts given (credits) and gifts received (debts/liabilities); "impure" means that these are gifts with all kinds of strings attached. Andy can accept a "gift" of bananas from Betty, but he has to give something of similar value (when the system works, the value is usually similar) to someone -- not necessarily Betty -- later (no bilateral trade balance; only multilateral, over time).

In my third post I put it like this:

"The central bank accountant follows two simple rules which he applies on a person-by-person basis: (1) if a person sells something, then credit her account, and (2) if a person buys something, then debit her account.

What is being recorded is how much each person has given or taken, without paying attention to who happened to be the counterparty in any particular trade. We know that Andy has taken goods worth SK100; from whom, it doesn't matter. We also know that Betty has given goods worth SK100; to whom, it doesn't matter. If we could trust that all transactions are legitimate, we could severe the link between a particular credit (entry) and a particular debit (entry), as long as the total credits equaled total debits on a given day (assuming records were updated only once, in the end of the day)."

JKH said: "but in fact the cheque instructs a green money transfer in the other direction (pay to the order of ... etc.)"

I don't think you can invoke conventional, or even legal, language to prove your point in this case. People and (thus) the law once said that the Sun revolves around the Earth, right? We can take Nick as saying that the Earth actually revolves around the Sun ;-)

JKH: Regarding your explanation, and alternative view, on TARGET2... I need to think hard if I have something intelligent to add to it :-) Very good. Perhaps I could try to give a new angle on it, but it might be time wasted. Perhaps we come back to the subject later on!

"JKH: In a red/green world, I think we can think about it either way."


You call a green money overdraft, red money.

So presumably, you would call a red money overdraft, green money?


I think I can invoke facts and obvious legality to prove my point.

The check says on the front "pay to the order of"

That's quite clear and quite legal.

If we then make the assumption that Nick's green/red framing of the real word is reasonable, the only conclusion to draw is that a real world cheque is a payment instrument that transfers green money balances in the specified direction of "pay to".

Think about it - buyer, seller, pay to the order of

It can't be otherwise logically or legally.


the transfer by cheque of a green money balance from a starting point of 0 results in an overdraft or a red money balance equal to the size of the transfer

but its still a green money balance rather than a red money balance that's being transferred

thinking about something in two different ways may include the way that it actually happens, and the way that it actually doesn't happen but hypothetically could happen with some appropriate modification

How about this situation.

I have a green money overdraft. I can clearly settle the overdraft by giving the bank a green money cheque from another green money account.

However, could it be extinguished by the bank paying me a red money cheque?

Is my green money overdraft reduced or does the bank raise another red money account in its name which is increased?

In which case there are now two accounts, each of different colour to the other, where previously there was one.

It's all getting impractical (costly presumably) and bordering on the absurd.

Living in a green/red money world could get complicated and confusing.

And presumably we can now call red money overdrafts green money.


“Nick, you call a green money overdraft, red money.

So presumably, you would call a red money overdraft, green money?”

The plot thickens.

I think that’s right.

A green money overdraft is created by spending a positive green balance down through the zero lower bound of green money – so creating red money as a result.

I think a red money overdraft would be created by spending a negative red balance up through the zero upper bound of red money – so creating green money as a result.

(“Spending a negative red balance” means, similar to green, getting rid of that balance. In the case of red money, that’s done by selling stuff, rather than by buying stuff as is the case with green money.)

"It's all getting impractical (costly presumably) and bordering on the absurd."

Either we crossed that bridge a long time ago, or we never will cross it.

"Either we crossed that bridge a long time ago, or we never will cross it."

LOL!!! You know what I am going to say to that one, don't you?

"In which case there are now two accounts, each of different colour to the other, where previously there was one."

And if this is correct, the cost of a dual green/red money system would be higher (account balances might be negated by complementary green/red transactions, but would grow ad infinitum).

And the higher cost might explain why a dual system has not naturally arisen.

Henry & JKH:

I wrote this on "red money overdrafts" to JKH in the other thread 1-2 days ago:

"We, of course, need to introduce overdraft into red-only world. This is no ordinary overdraft, but it's symmetrical opposite. It doesn't involve any credit. If you have an overdraft in the real or mixed world, you can buy without ever having a positive balance on your checking account. If you have an overdraft in the red-only world, you can sell without ever having a negative balance on your checking account -- that is, without possessing "red money" prior to the sale.

Nick might protest by saying that a positive balance on a checking account is not allowed in red-only world. But here you, JKH, and I should agree: a positive balance should be allowed, because that balance is not "green money" in red-only world. It's a non-monetary asset, just like a negative balance is a non-monetary liability in the real world."

JKH said: "e.g.

the transfer by cheque of a green money balance from a starting point of 0 results in an overdraft or a red money balance equal to the size of the transfer

but its still a green money balance rather than a red money balance that's being transferred"

Are we now talking about the real world? If yes, good. My point is that one cannot transfer balances. It's all arithmetics. One makes entries and those entries change balances. Already Schumpeter thought of it like that.

It sounds silly to think of transferring a credit ("green") balance when account balance is zero. Make a debit entry and you get a debit balance. Nothing transferred. Change one balance here by making a debit entry, and another balance there by making a credit entry. You don't need to transfer any balances; you don't need to decide whether its a credit or a debit balance you want to move, because you just cannot move an account balance. It's impossible.

You talked about two ways to view this. You're aware that I'm offering a third way?


I am a lazy old sod and not given to thinking too deeply about things unless I think there is something serious at stake.

Reading your second paragraph above leaves my head in a whirl - and I can't quite bring myself to trudge through the argument. I am over it!. :-) I'll leave it to JKH who is much more perspicacious than I am.

As I suggested a few days ago, Nick has created a monster and I am more than happy to leave the beast slumbering in the corner as I fear continual prodding might might stir him unnecessarily.

Henry:"unless I think there is something serious at stake"

How about the first real revolution in macroeconomics since Keynes? ;-)

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad