SAFE Policy Blog, 27. Februar 2019
Based on institutional specifications and accounting identities (rather than a model, as Krahnen maintains) our paper shows that Target balances measure a potential risk shifting between euro countries even if joint liability in monetary policy operations is formally excluded and no country leaves the Eurozone. The risk we assume is a loss in a national central bank's assets of any kind, including its assets from monetary policy operations such as providing refinancing credit to banks or buying marketable assets with fresh money.
Jan Pieter Krahnen argues that no relevant risks can occur due to a superior banking regulation under the Basel accord. The Basel accord in his opinion keeps the asset values of banks even in cases of huge losses – those which would only occur with a probability of 1% or even 0.1% - above the level of secured senior debt, a subset of which is central bank claims. The ECB's claims on banks are over-collateralized and safe. The losses, if any, "are borne by by private investors, not by taxpayers, and particularly not by central banks."
We do not endorse this argument for two reasons.
Firstly, we doubt that the models used in banking regulation to which Krahnen refers are appropriate to assess systemically correlated risks of the sort that shook the world in 2007 and 2008 and also in previous banking crises during earlier decades and centuries as reported by Reinhart and Rogoff (This Time is Different, Princeton University Press 2011) . These risks have little in common with the risks resulting from idiosynchratic changes in market values of assets as measured by the regulatory models used. These models are typically based on a support range of five years or less, which nowadays does not even include the Lehman crisis. (Are they "toy models" to use Krahnen's own words?) We are more than surprised about the confidence Jan Krahnen seems to have in the model-based approach to banking regulation that was shown to be utterly insufficient to assess the risks that culminated in 2007 and 2008 and unable to prevent Wall Street from turning into a gambling casino (Sinn, Casino Capitalism, Oxford University Press 2010). It is courageous assuming that this approach would in the future protect the ECB from bearing any risk from its refinancing operations. Did we not learn a lesson from the financial crisis?
If Krahnen's optimistic view of banking regulation was correct: Why then did the ECB council in 2012 decide to force the Greek central bank to convert all outstanding refinancing credits provided to commercial banks into ELA credit to avoid risk mutualisation due to an obvious under-collateralization. And why is it that authors like Admati and Hellwig (The Bankers' New Clothes, Princeton University Press 2013) keep demanding that banks hold 30% equity in the balance sheet (leverage ratio) rather than 3%, which is the minimum under the Basel agreement. And why did the EU countries find it necessary to give the permanent rescue fund ESM the task of serving as a back-stop for national bank resolution funds if all the risk is already borne by private share holders and private holders of senior bank debt?
Lending to banks is not even the most important source of Target risks
Secondly, Krahnen overlooks that claims on banks are only a minor part of the ECB's monetary assets. Most of the Eurozone's existing central bank money came into circulation not by providing refinancing credit to banks but by buying marketable assets issued by non-bank institutions, as we explained in our paper. Even an ultra-tough banking regulation cannot protect the Eurosystem against the risk of default in these assets. This is a trivial, but important point that the author seems to have overlooked.
Perhaps Krahnen misunderstood our term "money creation credit" and equated it with "refinancing credit". However, on page 37 of our paper we explicitly define what we mean:
“In this context, we refer to money creation credit as the sum of all measures that put central bank money into circulation in the broadest sense, i.e. not only refinancing loans of the usual kind, but also purchases of securities, including purchases under the PSPP programme and under the ANFA agreement. We also include ELA loans. The sum of the money creation loans is therefore equal to the monetary base (M0).”
We are sorry if this term caused confusion, but we needed it because the distinction between the various kinds of monetary assets in the central bank balance sheet was useless for our purposes and would have made the text clumsier.
The inclusion of asset purchases in the risk assessment is important insofar as most of the assets in the central bank balance sheets are nowadays government bonds and will likely stay that for a long time to come. However, government bonds are risky assets given that most Eurozone states are way above the Maastricht limit of 60% of GDP and as all versions of the Stability and Growth Pact that we have seen since 1996 have been persistently violated. Even if banking regulation were sufficient and thus would ensure safe banks, the regulation of government debt has not been respected at all, leaving a bunch of heavily over-indebted states in the Eurozone. Via the Target system, some of the government debt has effectively been mutualized in Europe even in cases where a mutualization was formally excluded.
Recall that the Eurozone has already experienced a number of incidences of sovereign bankruptcy in recent years, all applying to Greece. In early 2012, there was a haircut of 105 billion euros on Greek government debt at the expense of private shareholders. In the autumn of that same year, a debt conversion with an extension of maturities and an interest free period of ten years was introduced that cost public creditors 43 billion euros in present value terms. At the peak of the Greek crisis in Summer 2015 the directorate of the ESM formally stated another bankruptcy of Greece which then induced European governments to come up with a third rescue program. In 2018, this program was extended with maturities stretching way into the second half of this century and another period of ten years during which no interest was due. This was a hidden bankruptcy so as not to excessively irritate markets.
The possibility of sovereign bankruptcy also dwarfs the hope that national governments would be able to assume the debt of their central banks if the latter cannot fulfil their payment obligations inside the Eurosystem and hence become insolvent. If a central bank defaults because its government bonds default, any sort of legal obligation for a member state to recapitalize its central bank, even if lawyers would be able to read it out of the Maastricht Treaty or the ECB's bylaws, would be meaningless. The protection by a state whose insolvency triggers the insolvency of a national central bank is useless for the creditors of this national central bank.
Insecure government debt is not only a direct risk of central banks, but also an indirect risk because it may destabilize banks and hence impose a risk on the refinancing credit they received from the central banks. In the Basel system, government bonds receive a risk weight of zero in the calculation of risk-weighted assets. As is well-known this implies that the Tier-1 equity ratio which plays such a big role in banking regulation is basically a faked number, and the entire system of regulatory risk assessments based on that number is chronically and fundamentally false. We are not the only ones who come to this conclusion, which can nowadays be considered a mainstream opinion among economists.
A Public Choice view of the trustworthiness of the Basel System
Jan Pieter Krahnen argues that if, contrary to his own assessment, banks are not fully safe, regulators should make them safer rather than try to reduce the Target balances by more direct means. We disagree not only because Target risks resulting from purchases of non-bank assets by a national central bank, in particular those of the local state, cannot be reduced that way as explained above, but also because we distrust the forces that became operative in shaping the Basel agreement. All too often, banking lobbies have a significant influence on regulatory standards. Thus even if, theoretically, some of the Target risk could be reduced by resorting to tougher regulation we believe it is wishful thinking that a sufficiently tough regulation will ever come about. If public choice theory has told us a lesson, it is that we must not be naive in this respect.
As we explained in our paper, some of the Target balances came about, because private investors mistrust particular countries and therefore are not willing to provide credit to risk-prone national banking systems and governments at the same conditions at which the ECB is willing to do so. In other words, the ECB systematically undercuts local market conditions with cheap credit from the local printing press. This fact in itself is already a strong indication that markets do not trust the regulation of government borrowing and commercial banks by the originators of the Basel rules and the respective EU bodies, for if markets believed in the superiority of regulation, there would be no reason for them not to provide credit to endangered countries at the ECB's conditions, and the Target balances would not have arisen in the first place. Thus, in a sense, the Target balances themselves may show that markets do not seem to believe the reassuring story Jan Pieter Krahnen is telling us.
Distrust is not the only reason for the Target balances. The abundance of liquidity stemming from the QE program also played its role in recent years. (For the role of QE see H.-W Sinn, Der Schwarze Juni, Hanser 2016) But it was the dominant explanation during the first wave of rising Target balances until 2012, and this is the case to which our paper refers explicitly.
The blackmailing potential
There is one sense in which the Target balances might one day no longer indicate a particular risk shifting between Eurozone countries. As they impose a threat on Target creditor countries they can be used to pressure them to agree to fiscal transfers, insurance systems or debt mutualisation (Eurobonds, EDIS, ESBies, renewed OMT etc.). If enough of these schemes are operative the Target risks will have largely disappeared, but of course this is only because the tax payers of the creditor countries bear the risks through other channels.
In our opinion, this would be a highly problematic development because of the moral hazard and Dutch Disease effects resulting from badly designed transfer and mutualisation schemes. Europe needs more genuine private capital flows and private risk sharing which comes about without public intervention and is attracted by sufficiently high local risk premia in interest rates. Measures to contain the Target balances more directly would not only result in more prudent, safe and profitable investment activities, but would also reduce the blackmailing potential of Target balances by which the ECB has already put European tax payers on the hook.
Read on www.safe-frankfurt.de