Why the taxpayer is on the hook

Hans-Werner Sinn 24 February 2015



Would the taxpayers have to foot the bill?

It has repeatedly been attempted in the media to demonstrate that taxpayers would not have to foot the bill should losses occur from the purchases of government bonds by the ECB. The ECB balance sheet, it is asserted, is pure fiction. It can conjure up money out of thin air. German populists are stalking fear among the populace because they don’t understand this.

Is the German Constitutional Court then wrong to see a liability risk for the taxpayers in the unlimited purchases of government bonds of the crisis-afflicted countries under the SMP and OMT programmes? And did the opponents of the ECB’s QE programme fight in vain for leaving out of joint liability 80% of the purchases? Has the ECB given in to ‘German irrationality’?

Far from it. The Eurosystem is, save for the allocation of voting power (ECB 2015), arranged like a joint-stock company belonging to the Eurozone member countries. It is, of course, not really such a company, and its task is not to generate profits, but to carry out monetary policy. Nevertheless, as a by-product of its tasks, the system normally does generate profits, which are then distributed via the national central banks to the respective national treasuries. The profits result from the investment of the Eurosystem’s equity capital as well as from lending self-created money against interest, or from the acquisition of interest-bearing assets with such money. If central banks were traded in the stock exchange, they would have a market or present value that is derived, as it is for every other stock corporation, from their future dividend distributions.

Given the time path of the monetary base, any write-offs of defaulting government bonds lead to a permanent reduction in the dividends distributed to the member countries and to a decrease in the present value of these dividends in the exact amount of the write-offs. This is true regardless of whether or not accounting capital becomes negative in the process. The ‘shareholder countries’ must therefore either raise taxes or reduce expenditures. The assertion that this poses no problem since ‘merely’ the profits sink, as is sometimes maintained, is a euphemism. What would a real shareholder say if merely part of his dividends were taken away?

The loss of profit and taxpayers

It is correct, however, that the taxpayers are under no obligation to recapitalise the central bank in order to make up for the loss of profit (European Communities – Council 1992). But this holds for every real stock company as well, without this meaning that the shareholders would not perceive the drop in dividends as a loss; they will, of course, have to bear the profit drop fully.

But could we be talking about peanuts here? Unfortunately not. In static terms, that is, when the monetary base remains constant over time, the present value of the ECB system’s dividends amounts to the sum of the stock of central bank money and the central banks’ equity capital, irrespective of when and how quickly the currently very low interest rates return to normal levels. (After all, the present value of the interest revenue earned by a deposit is always equal to today’s cash value of the deposit if the actual time path of the deposit’s interest rate is used for the calculation of the present value.) By the end of 2014, the monetary base amounted to €1,317 billion, while the equity capital including valuation reserves added up to €425 billion. The assets to be distributed amounted thus to €1,742 billion. One can deduct from this sum the banks’ minimum reserves for which the ECB may, but doesn’t have to, pay interest. This would bring the sum to €1,636 billion. This is more or less equivalent to what German reunification has cost in terms of net transfers through the government budget. Assuming a continuous expansion in the monetary base on par with economic growth, one would even come to a present value of about €3.4 trillion, as Buiter and Rahbari (2012) have calculated. This sum can whet some appetites.

France, for example, is entitled to around one-fifth of this sum, i.e. almost €700 billion. It must also bear a fifth of the potential write-off losses resulting from government bond purchases already performed under the SMP programme or promised under the OMT programme, both programmes set up to prop up the Eurozone’s crisis countries. And it would also be liable for potential write off-losses of one-fifth of the 20% portion of the government bond purchases under the new QE programme which are to be pooled among the member countries.

Of course, in net terms, such losses would only result from other governments’ bonds purchased under these programmes. If, as is the case for 80% of the QE purchases of government bonds, a central bank buys its own government’s bonds, a default is not a net loss, since the disadvantage of the write-off loss is matched by the advantage of getting rid of a payment obligation of equal size (De Grauwe and Ji 2015).  

Thus, much to the chagrin of those claiming otherwise, there is no such thing as a free lunch for the participants of government bond purchase programmes that involve the international mutualisation of interest revenues (see also Sinn 2014).


Buiter, W and E Rahbari (2012), “Looking into the Deep Pockets of the ECB”, Citi Economics, Global Economics View, 27 February.

De Grauwe P and Y Ji (2015), “Quantitative Easing in the Eurozone: It´s Possible without Fiscal Transfers”, VoxEU.org, 15 January.

ECB (2015), "Organisation, Decision-making, Governing Council", FAQ on the rotation of voting rights in the Governing Council, 1 January.

European Communities – Council (1992), “Treaty on European Union”, in particular protocols, chapter 5, article 33, paragraph 2.

Sinn, H-W (2014), The Euro Trap. On Bursting Bubbles, Budgets, and Beliefs, Oxford University Press, Oxford, chapter 8: "Stumbling Along – No Risk for Taxpayers?", pp. 265-267.



Topics:  EU policies Taxation

Tags:  Eurozone taxpayers, profit loss, government bonds purchases


Paul De Grauwe and Yuemei Ji

Hans-Werner Sinn compares the central bank to a private company. This is an unhelpful comparison. A private company needs positive equity. If its equity turns negative, i.e. if the value of its liabilities exceeds the value of its assets, the company is insolvent.

This is not the case with a modern central bank, which can have negative equity without being insolvent. The reason is that the liabilities of the central bank (money base) do not create a claim on the assets of the central bank. The value of money issued by the central bank is independent of the assets the central bank is holding. Even if the central bank has no assets the money it is issuing (e.g. “helicopter money”) has a value, which depends on its purchasing power, i.e. the amount of goods and services it can buy. That is the nature of fiat money.

The comparison of the central bank with a private company is not only unhelpful; it is also misleading. Let us take the case of a central bank in a single country (we consider the case of a central bank in a monetary union later). When the central bank buys government bonds, it substitutes interest-bearing assets for monetary liabilities earning no interest and carrying an inflation risk. As a result, it changes the budget constraint of the government. By monetizing government debt the central bank relieves the government from paying interest on the debt. This is the moment seigniorage is created.

Most central banks maintain the fiction that the government bonds still exist economically by keeping them on their balance sheet. These bonds, however, are just a claim of one branch of the public sector (the central bank) against another branch of the public sector (the government). These two branches should be consolidated into the public sector and then it turns out that these claims and liabilities cancel out.

Another way to see the same thing is by considering the flows. When the government bonds are kept on the balance sheet of the central bank, the government transfers interest to the central bank. The latter then transfers this interest revenue back to the government.

It follows that the central bank could write-off these bonds (and reduce the equity in the same amount, so as to make the accountants happy). This would make no difference for the government budget (and the taxpayer). As long as the money base that has been created is not reduced, the government continues to enjoy the seigniorage (“dividends” in the parlance of Hans-Werner Sinn) that arises from the fact that it does not have to pay interest on the debt that the central bank has purchased. That seigniorage is independent of the value given to the bonds in the balance sheet of the central bank. It is therefore incorrect to claim, as Hans-Werner Sinn does, that “any write-offs of defaulting government bonds lead to a permanent reduction in the dividends” distributed to governments.

Things are a little more complicated in a monetary union where one central bank faces 19 governments, as is the case in the Eurozone. The complication arises because a bond-buying program can lead to a fiscal transfer between member-countries. In our article (De Grauwe and Ji(2015)) we show, however, that a bond-buying program can be organized in such a way that it does not lead to fiscal transfers even if one of the participating member countries’ government defaults. This can be achieved by making sure that the interest on the government bonds purchased from the different member countries is returned to the same governments. Such a rule of “juste retour” ensures that any write-off of government bonds due for example to a default, leaves the taxpayers of the other countries unaffected. Suppose, for example that the Italian government were to default. In that case it would stop paying interest to the ECB, which would then stop paying back the same amount of interest to the Banca d’Italia (and hence to the Italian Treasury). The other interest flows between the ECB and the other national central banks would be unaffected. There would be no fiscal transfer.

The recently announced ECB bond-buying program has this “juste retour” feature for 80% of the program. Technically this is achieved by keeping the government bonds on the balance sheets of the national central banks. These could also have been kept on the ECB’s balance sheet using the rule of “juste retour” described earlier.

20% of the bonds will be held by the ECB, creating a potential fiscal transfer between the member countries. How large is this potential transfer? The maximum exposure of Germany can be computed as follows. Under the QE arrangement, announced by the ECB, €54 billion of German bonds will be in the common pool of 20%. The average interest rate on German government bonds is 1.7%, i.e. the German Treasury yearly interest payments in the pool amount to €918 million. Assume now the worst possible scenario: all governments default on their debt, except Germany. This surely would mean a collapse of the Eurozone. But let’s use this as a benchmark to compute the maximum maximorum of German losses. In this scenario Germany would be the only one paying interest in the pool. This would be distributed to all member countries according to the equity shares. Germany would receive €248 million. Its loss would be €670 million (918 – 248). This represents 0.017% of German GDP. Thus even under the most extreme scenario the exposure of the German taxpayer to QE losses is peanuts.

Hans-Werner Sinn

After buying 223 billion euros worth of government bonds from crisis-afflicted countries under the Securities Markets Programme (SMP) and announcing unlimited purchases from such countries under the Outright Monetary Transactions programme (OMT), the ECB recently announced its QE programme, which entails purchases of more than 1.1 trillion euros worth of predominantly public assets. Eighty percent of the government bond purchases under the QE programme will be under the juste retour rule, and 20% will be carried out under a debt mutualisation rule.

Paul De Grauwe and Yuemei Ji emphasize that an outright government default under the QE programme for those bonds which are bought under the juste retour rule would have no redistributive implications. There is no disagreement about this. This is why I emphasised in my piece that my arguments only apply to the SMP and OMT programmes, which involve debt mutualisation, and that it was important to limit the mutualisation to 20% of QE. The authors are tilting at windmills.

A similar remark applies when they repeat the truth that central banks can operate with negative equity capital. The reader will realise that I said as much when I referred to the potentially negative accounting capital. I used the term accounting capital because the monetary base that the central banks lend out against interest or use to buy interest-bearing assets can be seen as an economic equity capital that must be added to the accounting equity capital, given that, except for the minimum reserves, the central bank does not pay interest on the money it issues. As is well known, the monetary base can even be seen as true economic wealth in net terms, because it delivers liquidity services to the economy.

So the only relevant point the authors make refers to the magnitude of the potential redistribution losses should some of the government bonds bought under mutualisation rule default. They refer to the potential loss in current interest flows for the 20% of QE that involve debt mutualisation, showing that it would be minuscule. I doubt that that is a valid counter-argument. For one, they again reiterate my point that only the debt mutualisation part of QE involves risks for others. For another, they overlook the fact that the ECB currently resorts to a zero-interest policy. With the same rationale one could argue that the owner of a savings account should not care about being deprived of his savings, given that he nowadays does not earn any interest anyway.

The argument obviously is wrong from an economic perspective, as interest rates will not be zero forever. Trivially, a defaulting portfolio will result in the loss of a permanent interest flow whose present value is the same as the current write-off losses themselves. This statement is true, even if the rate of interest is temporarily zero because of the current ECB policy.

Thus a euro country’s share in the potential write-off losses on defaulting government bonds bought by the ECB under one of its debt mutualisation programmes would be equal to the present value of this country’s share in the resulting permanent loss of interest income. For example, a country like France, with a share of 20% in the ECB’s profit distributions, would shoulder 20% of the losses on the 94 billion euros worth of Italian government bonds bought under the SMP programme, or up to 19 billion euros, should Italy default.

One might argue that France would not lose the 19 billion euros, because without the SMP programme there would not be an interest gain in the first place. But this is comparing apples and oranges, since a meaningful comparison can only be made with a given time path of the monetary base, because otherwise redistributive effects involving money holders affected by increased inflation or reduced deflation would have to be taken into account. Thus, the 94 billion euro SMP purchases of Italian government bonds could be compared, for example, with a 94 billion euro QE purchase under the juste retour rule involving all Eurozone government bonds. Assuming that no new government bonds will be issued because of the ECB’s secondary market purchases, the French government would receive an additional seignorage income flow under the QE programme whose present value is 20% of 94 billion euros, regardless of whether or not Italy defaults, but would receive nothing if Italy defaults on its entire debt under the SMP programme. The reader will be able to construct similar meaningful comparisons between QE programmes with and without a juste retour rule.

This shows how important it was to move from the old SMP to the QE programme with a juste retour rule, and that negotiating such a rule was not immaterial or irrational as has been claimed.

Professor of Economics and Public Finance, University of Munich; President, CESifo

Vox Talks