By now most readers of European financial newspapers and blogs will have come across Hans-Werner Sinn’s repeated assertions (see Sinn 2011a and 2011b on this site and his latest here) about Germany’s “stealth bailout” of European peripheral economies and the “ticking time bomb” hidden in the €300 billion claims of the Bundesbank in the Eurozone payment system TARGET2. Journalists (in English e.g. Storbeck 2011a, Alloway 2011 and Smith 2011), academics and central bankers at VoxEU (Whelan 2011) and elsewhere (e.g. Bindseil and König 2011a, Buiter et al. 2011, Bundesbank 2011) have since convincingly refuted Sinn’s arguments.
The origin and meaning of such imbalances between central banks, however, has never been comprehensively explained in the debate, which might be a reason why Prof. Sinn’s theses continue to circulate (see Paul Krugman 2011 and Martin Wolf 2011). Here, I propose to take a step back and think more systematically about how the balance sheet(s) of the Eurosystem banks interact. We will see:
- that large imbalances can (and do) arise even without current-account deficits (Sinn) or banking crises (his critics) just because of the normal functioning of Eurozone,
- that banknotes have to be included in the analysis (which is done here for the first time),
- that (somewhat ironically) the Bundesbank had considerable debts within the Eurosystem before 2007,
- therefore Prof. Sinn’s and Prof. Wollmershäuser’s recommendations to limit TARGET imbalances are not only impractical but are incompatible with a monetary union.
A very simple central-bank balance sheet
Let’s start with a very simple central-bank balance sheet (Figure 1). The central bank issues money in the form of banknotes and deposits held by commercial banks (mainly in order to fulfil minimum reserve requirements). These are on the liability side. The central bank issues the money either through lending or by buying assets outright. To simplify the presentation, in Figure 1 both items are on a net basis. In the case of lending to banks this means subtracting liquidity absorbing operations from liquidity providing transactions.1 For net outright assets we add holdings of gold and foreign exchange and subtract liabilities like capital and reserves, among others. As can be seen in Figure 2, since the inception of monetary union the Eurosystem balance sheet has been mainly driven by an increase in banknotes, matched by an increase in lending operations on the asset side.
Figure 1. The most simple Eurosystem balance sheet
Figure 2. The development of the Eurosystem balance sheet since 2001
Enter Eurozone, TARGET and other intra-system claims
So far so good. In the Eurozone things get a little bit more complicated because monetary policy is decided centrally but implemented decentralised through the 17 national central banks. If a Spanish bank borrows money from the Eurosystem, the operation will show up on the balance sheet of the Banco d’Espana, while an operation with a French bank will appear at the Banque de France. Bank deposits as well as outright holdings are equally reported in the balance sheet of the respective national central bank. What about banknotes? Banknotes do not show up at the central bank that physically put the note into circulation, but are divided up among all national central banks according to a key that reflects the countries’ economic weight. This is necessary, because the assets received in return for banknotes issued generate seignorage that has to be distributed fairly within the Eurosystem (for a comprehensive discussion see Handig and Holzfeind 2007).
The consequence of these financial reporting rules of the Eurosystem is that while the accounting identity (net lending + net outright assets = banknotes + deposits) holds on the consolidated Eurosystem level, this is not the case for the individual central bank. The difference between assets and liabilities will show up as a claim or a liability towards the Eurosystem. These are the intra-Eurosystem balances, which the infamous TARGET2 accounts are a part of. Using this simple balance sheet framework we can now think about various scenarios that lead to changes in these intra-Eurosystem balances.
Financial centre or financially stressed?
In a monetary union there is no reason why banks in country A will obtain reserves and banknotes from their central bank (A-CB). Instead they could borrow in the interbank market from banks in country B that in turn borrow at their central bank (B-CB). Figures 3 and 4 show the effect on the national central banks’ balance sheets. In Figure 3 the B banks increase their net borrowing and lend the money in the interbank money to the A banks. The transfer is done through TARGET and B-CB incurs a TARGET liability. On the receiving side, the A-CB reduces net lending to its banks (because they borrow from the B banks). The share in total banknotes however remains the same and so do minimum reserve requirements. In the case of A-CB, the balance is made up by a TARGET claim. The TARGET balances thus serve to equalise assets and liabilities of both national central banks.
Figure 4 shows an alternative scenario where the transfer is not done through TARGET but by shipping banknotes. This is not as absurd as it sounds – as we will see in a minute, German banks traditionally provide banks in other euro countries as well as outside the Eurozone with banknotes. The effect on the balance sheet is similar with the only difference that these transfers show up not as TARGET claims/liabilities but as “net liabilities/assets related to the allocation of euro banknotes within the Eurosystem”.
Figure 3. Effect on national central bank balance sheets part 1
Figure 4. Effect on national central bank balance sheets part 2
The first important (and fun) result is that the same balance sheet patterns will arise in the case of a run on banking system in country B. Figure 4 is a traditional bank run – savers want cash. Figure 3 is a wholesale capital flight – the A banks require repayment of interbank loans, the B banks get the funds from their national central bank and pay through TARGET (how interbank payments result in TARGET has been discussed extensively, see Bindseil and König 2011). Just by looking at intra-system balances we cannot distinguish a bank run from a tiering of banks within monetary union. And this is the second important result. There is no material difference between “TARGET claims/liabilities” and “claims/liabilities related to the allocation of banknotes”.
Is the Bundesbank lending to the GIPS or is everybody lending to the Bundesbank?
The choice of the scenarios was of course not arbitrary, as will be seen by looking at the balance sheet of the Bundesbank in Figure 5. Again we see a trend growth of banknotes and monetary policy operations. In addition, between 2002 and 2007 the Bundesbank built up liabilities to the Eurosystem of about €70 billion that have since turned into significant claims of about €160 billion. Why? The answer is given in Figure 6, which gives the share of the Bundesbank in the various components of the Eurosystem balance sheet. Banknotes stagnate close to 30%, which they have to because banknotes are divided up according to the capital share of each national central bank in the Eurosystem. Bank deposits are also roughly constant around 30%. At the same time, however, before 2007 German banks borrowed heavily, more than 50% of total Eurosystem lending.
Figure 5. A simplified balance sheet of the Bundesbank
Figure 6. Bundesbank balance sheet in % of Eurosystem balance sheet
What have German banks done with the borrowed money? This will decide where the liability matching the “excessive” lending of the Bundesbank will show up in its balance sheet. The German banks can send it abroad electronically, thereby causing the Bundesbank to build up a TARGET liability (this is what Irish banks have done recently). Alternatively they can withdraw cash. As Figure 7 shows, German banks did just that, and the Bundesbank increased their liabilities from an “excessive” issue of banknotes. We cannot say with certainty what happened with these banknotes. But there are very strong indications that these banknotes were exported, partly to other Eurozone countries, partly outside the Eurozone (see Bundesbank 2011). After 2007 the situation reversed. As demand in Eurosystem operations came increasingly from banks in peripheral economies, demand by German banks declined. Net liabilities turned into net claims.
Figure 7. Claims and liabilities of the Bundesbank within the Eurosystem
Conclusion: Why restrictions on TARGET won’t work and would be harmful to Germany
By now it is hopefully becoming clear why the latest proposals by Prof. Sinn and Prof. Wollmershäuser won’t work and don’t make sense. First, restrictions on electronic transfers through TARGET could be easily circumvented by sending cash. Secondly, claims and liabilities within the Eurosystem will arise for various reasons, many of which are benign or even welcome, e.g. when banks within an integrated monetary union specialise and financial centres emerge. The sums involved are large. Annual settlement of intra-system debts (as in Prof. Sinn’s latest proposal) would have required the Bundesbank in 2006 to pay close to €80 billion, equivalent to its entire gold and foreign exchange holdings (€81 billion). Short of selling all its gold, the Bundesbank would have had to prevent German banks from playing an intermediating role in the Eurozone. This can neither be in Germany’s best interest nor consistent with the idea of an integrated monetary union.
Postscript: Differences with the US system
So how can the US Federal Reserve banks settle their intra-system claims every year? In addition to the point by Buiter et al. (2011) that claims can be settled by simple base money creation, there is also a structural difference between monetary policy implementation in the Eurozone and in the US. Operations in the US are mainly executed through a centralised account (System Open Market Account), not through the individual balance sheets of the reserve banks (equivalent to the national central banks in the Eurozone). SOMA-assets are allocated across reserve banks in rough proportion to the capital paid. That is why similar imbalances due to monetary policy operations cannot arise in the US.
Storbeck, Olaf (2011a), “The stealth bailout that doesn’t exist: debunking Hans-Werner Sinn”, Olafstorbeck.com, 6 June.
Storbeck, Olaf (2011b), “Schluss mit der Langeweile. Leseliste zur Target2-Debatte”, Handelsblatt Blog, 20 June.
Alloway, Tracy (2011), “The wonkiest web debate ever – Germany’s ‘stealth bailout’”, Financial Times, 8 June.
Bindseil, Ulrich and Philipp Johann König (2011), The economics of TARGET2 balances. SFB 649 Discussion Paper 2011-035, Humboldt University Berlin.
Bundesbank (2011a), Foreign demand for euro banknotes issued in Germany, in Monthly Report, January 2011.
Bundesbank (2011b), The dynamics of the Bundesbank’s TARGET2 balance, in Monthly Report, March 2011.
Buiter, Willem, Ebrahim Rahbari, and Jürgen Michels (2011), “TARGETing the wrong villain: Target2 and intra-Eurosystem imbalances in credit flows”, Citi Global Economics View, 9 June.
Handig, Martin and Robert Holzfeind (2007), “Euro banknotes in circulation and the allocation of monetary income within the Eurosystem”, Monetary Policy & Economy Q1 2007.
Jobst, Clemens (2009), “Monetary policy implementation during the crisis in 2007 to 2008“, Monetary Policy & Economy Q1 2009.
Krugman, Paul (2011), “The Euro Living Dangerously”, The New York Times, 1 June.
Sinn, Hans-Werner (2011a), “The ECB’s stealth bailout“, VoxEU.org, 1 June.
Sinn, Hans-Werner (2011b), “On and off target“, VoxEU.org, 14 June.
Sinn, Hans-Werner and Timo Wollmershäuser (2011c), “Target loans, current account balances and the ECB’s rescue facility“, CES-ifo Working Paper 3500.
Smith, Geoffrey T (2011), “Sorry, Professor Sinn, You're Way Off Target This Time”, The Wall Street Journal, 9 June.
Whelan, Karl (2011), “Professor Sinn misses the target”, VoxEU.org, 9 June.
Wolf, Martin (2011), “Intolerable choices for the eurozone”, Financial Times, 31 May.
Many items of central bank balance sheets are subject to strong daily fluctuations. End of year balance sheets can therefore be misleading on the normal size of items like bank deposits or monetary policy operations. This is why in this article annual averages are used. For the consolidated Eurosystem balance sheet these annual averages are calculated from the weekly statements see here. The Bundesbank makes its balance sheet available monthly in its monthly reports, see here. Unfortunately not all positions are detailed there. While the “Intra-Eurosystem liability related to the euro banknote issue” is provided, all kinds of intra-Eurosystem claims are subsumed under “Other assets”. However, we know from the annual reports that the other items besides intra-Eurosystem claims that are included under “Other assets” are small and relatively stable. Therefore it seemed safe to calculate the average “Net Target claims” used above by subtracting from the average “Other assets” from the monthly reports the relevant end of year items from the annual reports.
1 Normally, in the Eurozone liquidity absorbing operations are very rare. During the crisis large amounts of excess liquidity were absorbed through the deposit facility (see e.g. Jobst 2009).