Explaining ELA and the Consequences of Grexit on Target2
ELA works like this. Central banks will lend reserves to banks against a range of collateral, subject to haircuts and conditions. In the Eurosystem, provision of “emergency liquidity assistance” (ELA) is the responsibility of national central banks, though it requires ECB approval.
ELA is provided to Greek banks by the Hellenic Central Bank, which bears all the risks associated with it. Professor Sinn’s assertion that the risks of lending rebound to other central banks in the Eurosystem is flatly contradicted by the ECB:
ELA means the provision by a Eurosystem national central bank (NCB) of:
(a) central bank money and/or
(b) any other assistance that may lead to an increase in central bank money to a solvent financial institution, or group of solvent financial institutions, that is facing temporary liquidity problems, without such operation being part of the single monetary policy.Responsibility for the provision of ELA lies with the NCB(s) concerned. This means that any costs of, and the risks arising from, the provision of ELA are incurred by the relevant NCB.
In the event of Greek default, the Hellenic National Bank would become technically insolvent because of all the Greek sovereign debt that has been pledged to it by Greek banks. But it would be the responsibility of the Greek sovereign to recapitalise it, not other central banks.
If Greece defaulted and left the Euro, the Hellenic National Bank would acquire the right to create the new national currency – a right it does not currently possess. Greek government debt, we assume, would be redenominated in the new currency (lex monetae). Grexit, therefore, would resolve the Hellenic National Bank’s “insolvency”. However, that would create a problem. What would it do with the Euro-denominated reserves on its balance sheet?
The simple answer is that it would also convert those to drachma. Greek banks would then be unable to meet demands for Euro deposit account withdrawals. Greece would have no choice but to impose capital controls and a bank holiday to avoid bankrupting the banks.
But the conversion of both Euro-denominated reserves and their collateral (Greek sovereign debt) to drachma would have no effect whatsoever on other central banks. There would be no losses anywhere else in the Eurozone. Sinn is simply wrong about ELA.
But what about Target2?
The question of Target2 balances is somewhat more complex. Target2 is the Eurosystem’s real-time gross settlement (RTGS) system. All Western central banks have RTGS systems: they are the core of the electronic payments systems upon which Westerners have come to depend. Target2 is a little more complex than the RTGS of a single country such as the UK.
But only a little more complex. It is in reality far more straightforward than a lot of the rubbish that is written about it suggests.
Example 1: How asymmetric trade flows cause Target2 imbalances
The Bank of England’s RTGS system can settle payments between people in London and people in Manchester. If people in Manchester make lots of purchases from companies based in London, funds flow from Manchester to London. But we use double entry accounting to record all movements of funds. So a net flow of say £1m private sector funds from Manchester to London through the Bank of England’s RTGS system looks like this:
Manchester private sector DR cash at bank £1,000,000 (asset) DR e.g. fixed assets £1,000,000 (asset)
Manchester banks DR customer deposits £1,000,000 (liability) CR reserves £1,000,000 (asset)
London banks CR customer deposits £1,000,000 (liability) DR reserves £1,000,000 (asset)
London private sector CR cash at bank £1,000,000 (asset) CR inventory £1,000,000 (asset)
This is what is known as “quadruple accounting”, where double entries are recorded for all four participants. You can see that there has been a movement of goods (fixed assets) from London to Manchester. You can also see that there has been a movement of cash from banks in Manchester to banks in London, which is recorded both as a change in “cash at bank” on the customer side and as a change in “customer deposits” at the banks. This is NOT two lots of money: it is the same money, recorded as an asset & liability pair (customer asset, bank liability). You can also see that there has been a movement of reserves from banks in Manchester to banks in London (remember CR to a bank’s reserve account reduces the balance).
Now let’s add in the central bank’s reserve accounts. Remember that reserves are assets for commercial banks. The corresponding liabilities are held at the central bank. We can think of reserve accounts at the central bank as similar to a bank customer’s transaction account: it contains a small amount of moving funds. So when there is a net flow of funds from Manchester to London, the reserve movements look like this:
Bank reserve accounts (assets):
Manchester banks CR £1,000,000
London banks DR £1,000,000
Central bank reserve accounts (liabilities):
Manchester DR £1,000,000
London CR £1,000,000
So at the central bank there is a reserve imbalance between Manchester and London. Manchester is in “deficit” and London is in “surplus”. Or, putting it another way, Manchester has a net liability to the central bank and London has a net claim on it. As the entries balance, we could ignore the central bank completely and say that London has a net claim on Manchester.
But this is silly. London’s private sector has already received cash. All the reserve entries show is that there has been a net flow of funds from Manchester to London. In this case it is balanced by a net flow of goods in the other direction. What this is showing, therefore, is that London has a trade surplus and Manchester a trade deficit. In no sense does Manchester “owe” London anything. It has already paid.
Now, using the above example, replace Manchester with Greece, London with Germany, and the Bank of England with the ECB. All Target2 does is facilitate and record these movements of funds. It is a gross misunderstanding of RTGS settlement accounting to describe the imbalances arising from these movements as “debts”, as Professor Sinn does.
So we can see clearly how Target2 records the trade imbalance between Germany and Greece. But there was a large trade imbalance between Germany and Greece before the financial crisis. Why did this not show up as a Target2 imbalance?
The reason is that prior to the financial crisis, Greece’s trade deficit with Germany was funded by borrowing from German banks. Let me show you how this works.
Example 2: How foreign financing of trade deficits eliminates Target2 imbalances
Greek customer takes out a loan of 1,000,000m EUR from a German bank (in practice this was often mercantile credit, i.e. importer borrowed from exporter who in turn borrowed from his own local bank, but let’s not complicate things). The loan accounting entries are as follows:
Greek customer CR German bank loan 1,000,000 (liability) DR cash at German bank 1,000,000 (asset)
German bank DR Greek customer loan account 1,000,000 (asset) CR Greek customer deposit account 1,000,000 (liability)
The Greek customer pays the money to the German exporter. The accounting entries are as follows:
Greek customer DR goods & services received 1,000,000 (asset) CR cash at German bank 1,000,000 (asset)
German exporter CR inventory 1,000,000(asset) DR cash at bank 1,000,000 (asset)
German banks (aggregate) DR Greek customer deposit account 1,000,000 (liability)
CR German exporter deposit acct 1,000,000 (liability)
RTGS reserve accounts (liability):
DR Germany 1,000,000
CR Germany 1,000,000
In other words, although the customer is Greek, the financial transaction in effect takes place entirely within Germany. This is why there were no Target2 imbalances even though Greece had a large trade deficit with Germany.
When the Greek crisis hit, German banks stopped financing German exports to Greece. Greek customers were forced to borrow from Greek banks instead (you can see this clearly as a spike in Greek bank lending in early 2010 in the chart above). The result was a large Target2 imbalance.
But as the Greek economy faltered, imports to Greece fell massively. Trade stopped being the main cause of the Target2 imbalance. What replaced it was capital flight. And it is capital flight, not trade, that is currently causing the Target2 imbalance to grow.
Example 3: How capital flight exacerbates Target2 imbalances
Suppose we have a Greek who has £1mEUR of cash deposits at Greek banks. Fearing capital controls, redenomination and seizure of his deposits, he decides to move his money to safety in Germany. So he opens a deposit account at a German bank and transfers his money electronically from his Greek bank deposits to the German bank. The accounting entries look like this.
Greek customer CR Greek bank deposits 1,000,000 (asset)
DR German bank deposit 1,000,000 (asset)
Greek bank DR customer deposits 1,000,000 (liability)
CR reserves 1,000,000 (asset)
German bank CR customer deposits 1,000,000 (liability)
DR reserves 1,000,000 (liability)
RTGS reserve accounts (liability):
Greece DR 1,000,000
Germany CR 1,000,000
The Greek customer’s decision to move his money to safety widens the Target2 imbalance.
Because Professor Sinn believes that Target2 “deficits” are actual debts, and Greece is already very highly indebted, he thinks that Greece should take steps to stop the Target2 imbalances increasing. Greece should therefore impose capital controls so that our Greek customer can’t move his money to safety outside Greece. Presumably Professor Sinn also thinks that Greece should eliminate its trade deficit, though he doesn’t say this.
Central banks do not allow banks to run persistent reserve account deficits. Greek banks experiencing capital flight therefore have to borrow reserves to top up their reserve accounts. At present, because no-one else will lend to them, they borrow reserves from the Hellenic Central Bank, as I explained above. The Hellenic Central Bank’s balance sheet is therefore expanding by an amount corresponding to the growth of Greece’s Target2 deficit. But that does not mean they are the same thing. Professor Sinn unfortunately confuses them.
How would Grexit affect Target2?
Much time and energy has been spent discussing what the effect on Target2 – and, by extension, the ECB – would be if Greece left the Euro. Unfortunately most of the explanations are wrong.
Currently, Greece is running a Target2 deficit which is entirely covered, as far as Greek banks are concerned, by ELA from the Hellenic Central Bank. Suppose that Greece defaults, creates a new currency and redenominates all sovereign debt and all bank reserves into drachma, including reserves created through ELA. What does this do to Target2?
Nothing. Nothing at all. Zilch. Nada. Nix.
The Target2 “deficit” would remain as a notional liability of the newly-independent Greek state. It would still be denominated in Euros, since Greece would have no power to redenominate it. It would be frozen, since Greek capital controls and suspension of external trade in Euros would mean no further Target2 transactions. It would not need to be settled, paid, reallocated or otherwise disposed of. It could simply be ignored.
“But”, I hear you say, “surely there’s a catch?”
There is no catch. The existing Target2 deficit for Greece is entirely balanced by payments already made from Greek banks to banks elsewhere in the Eurozone. No-one is going to lose any money if Greece stops using Target2 because it ditches the Euro.
I admit, it has taken me quite a while to “get” this, despite the promptings of my good friend Beate Reszat who has always insisted that Target2 is simply a “black box” computer system and has nothing whatsoever to do with national accounting. But having now worked through the accounting, I am sure of my ground. The whole Target2 imbalances issue is a complete red herring.
If someone felt like being tidy, they could simply eliminate Greece’s notional Target2 deficit by proportionately reducing the Target2 surpluses of its Eurozone trade partners. In fact as the balancing reserves in Greek banks would already have been redenominated into drachma, this would technically be the correct thing to do. The Hellenic Central Bank leaving the Eurosystem and redenominating Greek bank reserves into drachma would reduce the aggregate quantity of Euro-denominated reserves in the Eurosystem. Bringing Target2 into line with this would preserve the consistency of Target2 balances with Eurosystem reserves. But it isn’t strictly necessary.
Nor is Greek redenomination potentially inflationary for the rest of the Eurozone, as some have suggested. Rather the reverse, actually. Capital flight can be inflationary for “safe haven” countries. If Greece left the Eurozone, capital flight would stop due to capital controls. This would be deflationary for the rest of the Eurozone, not inflationary.
Once the dust had settled, of course, Greece would want to lift capital controls and start trading with the Eurozone again. Most trade would probably be in Euros and settled via Target2. But that is true for any non-Eurozone country trading in Euros with the Eurozone. The Euro would be a foreign currency for Greece. It would have to earn euros through trade. And because of this, it would need to run a substantial trade surplus, helped by inevitable devaluation of the drachma. So if it were not “tidied up”, Greece’s Target2 deficit would shrink over time.