It came to be known as the ‘doom loop' – the mutually destructive relationship in which government bond-laden European banks weighed on the creditworthiness of their states, threatening the survival of both. But while the phenomenon has had a neat label for at least five years, it still lacks a neat solution.
That is not because there isn't one, but because any fix means tackling the status quo, whereby banks can hold domestic government bonds without a capital cost – a stance that makes little regulatory sense, but offers clear benefits to states with big borrowing needs.
If you zero-weight something, people will naturally hold too much of it
Paul Tucker, Systemic Risk Council
"If you zero-weight something, people will naturally hold too much of it. It is close to being a piece of social engineering, effectively subsidising credit to the state, making it a peculiar and distorting policy for regulators to pursue," says Paul Tucker, chair of the Systemic Risk Council and former deputy governor of the Bank of England, who stresses he is offering a personal opinion.
The obvious policy response is to ratchet up capital requirements or limit portfolio sizes, but either approach would hit some banks and states much harder than others – domestic government bonds make up just over 10% of the assets held by Italy's banks, for example (see figure 1). This introduces political strands into what might initially resemble a prudential debate and the knot is tightened by the special case of the eurozone, where member states cannot print their own money and therefore present a heightened default risk.
After years of collective frowning and chin stroking, recent months have finally seen support building for some new ideas. If a capital hit will not work on its own, for example, perhaps it would be better deployed in combination with some kind of concentration limit. Individually, each of these elements has its proponents, so a hybrid would offer something for everyone.
"The hybrid solution is probably the one on which it is easiest to find political agreement across all the countries, because everyone can say their own approach found its way into the final product. If banks go beyond the concentration limit, you could penalise with a uniform risk weight for all government debt, no matter which country," says Guntram Wolff, director of Brussels-based think-tank Bruegel and a former European Commission economist.
For now, the ball is in the court of the Basel Committee on Banking Supervision, which has indicated it will publish proposals on the treatment of sovereign credit exposures later this year or in 2017.
Existing Basel standards allow banks not to apply their models to sovereign credit risk exposures, but only for "non-significant business units and asset classes that are immaterial in terms of size and perceived risk profile".
The EU's fourth Capital Requirements Directive (CRD IV), which implements the Basel rules, took that a big step further by granting a broad, permanent exemption that allows banks to apply a standardised approach set by the domestic regulator – in many cases, national authorities have set a zero risk weight for exposure to the home sovereign. This earned the EU a rebuke from the Basel Committee in 2014, which found the rules were "materially non-compliant" with the international standard.
However, Europe is more dependent than the US on bank financing for the economy in general. So, while the Basel Committee is holding the rule-drafting pen, European authorities have been keen to have their say. In July, a meeting of EU finance ministers issued a statement supporting Basel's efforts to complete post-crisis reforms to capital requirements this year, but stressing the committee should be "taking into account in its global calibration also the distribution of its impact on the different banking models and across jurisdictions".
A paper published by the Dutch presidency of the Council of the EU in April drew heavily on a 228-page European Systemic Risk Board (ESRB) report on the regulatory treatment of sovereign exposures, published in March 2015, with an added dose of political calculation.
In its attempt to square the circle, the Dutch presidency offered five alternatives for EU governments to consider. The first was simply to preserve the status quo. The second was to enhance the treatment of sovereign risk under Pillar 2 and Pillar 3 of the capital rules, which relate to supervisory capital add-ons and public disclosure, respectively. The remaining three scenarios all involved changes to banks' Pillar 1 capital requirements.
Replacing the zero-weight exemption with modelled weights is one option. Fitch Ratings has examined those banks that do not use the CRD IV permanent exemption and applied the sovereign risk weights generated by their credit models, under Basel's internal ratings-based (IRB) approach, to the portfolios of 86 eurozone banks that do apply the exemption. These banks represented 92% of sovereign exposures in the European Banking Authority's (EBA) 2015 transparency exercise, which itself covered around 70% of EU banking assets. The results show a modest impact for most countries, but a dramatic effect for a few (see figure 2).
This may lead to accusations that the new regime was designed to deliberately punish banks in certain countries – a fear expressed by ESRB secretary-general Francesco Mazzaferro, speaking at the Institute of International Finance spring meeting in Madrid in May.
"From a completely personal view, I am worried this is perceived by the public as another case of north versus south in Europe. The idea in the south [is] that the north has a punitive attitude and wants to consciously create problems for banks in the south, which are full of sovereign [debt]. On the other hand, the north thinks banks in the south are not serious and are embarking on a strategy of gambling for resurrection by buying as many bonds as possible, so they are sure if there is any problem, they are going down together with their sovereign, so they will be rescued by somebody else," said Mazzaferro.
The idea of risk weighting may also fall down on technical grounds. In March, the Basel Committee issued a consultation that cast doubt on the viability of using IRB for low-default portfolios. Sovereign exposures were not included in the consultation due to the separate ongoing workstream, but the consultation implies the weight of opinion on the Basel Committee would oppose using internal models for sovereign debt.
Regulators could eschew models and instead opt for a new standardised approach that excludes the possibility of zero risk weights. The Basel Committee proposed a revised standardised approach to credit risk in December 2015, again leaving sovereign exposures for the separate workstream. The proposed standardised approach would rest heavily on the use of external credit ratings, but this runs counter to the broader desire among regulators to reduce mechanistic reliance on the rating agencies.
Measures based on credit ratings have a problem of potential procyclicality
Marco Pagano, Naples University
"Measures based on credit ratings have a problem of potential procyclicality, and the ESRB paper made one thing very clear: if you go back to what happened in the crisis when we had risk weights based on credit ratings, risk weights would have risen far too late, compared with the period when risk was going up, and they would have stayed too high for too long," says Marco Pagano, professor of economics at Naples University and one of the authors of the ESRB's 2015 report.
There is also a question about how effective raising risk weights would actually be in the current environment of ultra-low interest rates and wide disparities between the spreads of different eurozone sovereign bonds. Nishay Patel, head of European rates strategy at UBS, says German 10-year bunds were yielding around zero in mid-July, while the return on equivalent Spanish government bonds was 1.1%.
"If you really wanted to deter the type of carry trades that were set up by banks at that time [before the eurozone crisis], you would have needed risk weights well above 100%, which is almost unthinkable. How high do you have to push them to truly deter carry trades by banks in the context of a sovereign debt crisis? We have no idea, because we don't have the evidence – we don't have a track record to see what is the elasticity of banks' portfolio choices with respect to risk weights," says Pagano, who chairs the ESRB's advisory scientific committee.
A concentration limit, similar to the one that already restricts holdings of corporate debt, would avoid some of the difficulties associated with risk weights. It would precipitate a bond sell-off, but this would be a one-off adjustment, and it would certainly be effective in limiting bank holdings of sovereign debt.
But it has problems of its own. The current concentration limit for corporate exposures is 25% of capital, but this looks far too tight for sovereign exposures that are approaching 200% in many countries (see figure 3). Market participants tend to assume a 100% limit is a realistic starting point. Even on this basis, Fitch found banks would need to reallocate €492 billion in home sovereign exposures to avoid breaching the limit.
In April, Banca d'Italia published a paper on the prudential treatment of sovereign exposures, entitled ‘Easier said than done?' One former regulator suggests this could be an early indication of Italy's negotiating position in any eventual discussion on implementing sovereign exposure reforms.
Banca d'Italia concluded that introducing risk weights might be "manageable". Dependence on ratings agencies could be avoided by applying publicly available fiscal sustainability measures such as those used by the International Monetary Fund and European Commission. Conversely, the paper warned "imposing tight concentration limits on sovereign exposures could have significant effects", including a 51 basis point rise in government bond yields for a limit set at 100% of capital.
"The banks have very large holdings right now. The possibility of diversifying is not that great – the demand for credit is not very much – so it is desirable in the long run, but you need to introduce it carefully, sequence it, give everyone time, and you may need some form of public support," such as ECB intervention, says one capital markets banker.
One way of moderating the impact of a concentration limit is a hybrid approach, proposed as the fifth and final alternative in the Dutch presidency paper. Under this method, home sovereign exposures below the concentration limit would be subject to a low or zero risk weight, which would rise for any exposures over the limit.
This idea seems to be finding favour among regulators. In July, EBA chiefs Andrea Enria and Adam Farkas published an article advocating a 100% concentration limit, with "a gradual and increasingly steep increase in risk weight when concentration of exposures towards a sovereign rises above that threshold".
This approach would ease the risk of a sudden sell-off in sovereign debt, which could be caused by a hard exposure limit. It would also allow banks with a diversified sovereign portfolio to increase holdings and act as a stabilising influence during bouts of market volatility. Bruegel's Wolff suggests applying a uniform risk weight to avoid undue and asymmetric selling pressure on specific lower-rated sovereigns.
Replumb the market
Even the hybrid approach poses major challenges. In its research, Fitch Ratings assumed either IRB or a flat standardised risk weight of 10% for exposures below 100% of capital. Above the threshold, Fitch assumed the EU would use the concentration risk capital component multipliers already applied to large exposures in the trading book under the capital requirements regulation (see Table A).
If applied to all general government holdings over 100% of capital, this would lead to an increase of just under €800 billion in risk-weighted assets for eurozone banks in the sample, or a rise in capital requirements of almost 25%. The EU might well opt for a less severe risk-weight multiplier, but any meaningful hybrid limit would still be challenging.
"In practice, we would not expect the banks [to] simply accept these new capital charges and not rebalance or make adjustments to their portfolios. Many of them would adjust, and the precise extent of the adjustment would depend on the precise details of the implementation and how aggressively the excess charges scaled up, but it is not possible to know precisely what that effect is without more clarity on the absolute magnitude of the risk weights and the extent to which they scale up," says Alan Adkins, group credit officer for financial institutions at Fitch Ratings and a former Bank of England regulator.
In theory, banks in core eurozone countries needing to meet their own sovereign exposure limits might switch into peripheral European debt. This would ensure no overall drop in demand across the monetary union as a whole.
Jose Maria Roldan
"Home bias has to be eliminated, but home bias works in two ways: ring-fencing in case of crisis, and prohibiting your home institutions from purchasing foreign debts is another form of home bias that has been very damaging during the crisis," says José María Roldán, chief executive of the Spanish Banking Association and a former director-general of regulation at the Banco de España.
The exact path of diversification is uncertain, however, and the transition could cause market volatility and disruption. Fitch estimates as much as 10% of Dutch general government debt, and 8% of Spanish and German general government debt, could be affected by a 100% limit (see figure 4). Moreover, the end result would still carry systemic risks.
"If we manage to push all the banks to hold well-diversified portfolios of sovereign debt from across the eurozone, this is a very good thing from the point of view of diversification, but it may potentially be very problematic from the point of view of systemic risk, because everyone would have similar portfolios and there could be contagion: if the countries of the periphery were in trouble again, the banks in the core eurozone might at that point also have to be recapitalised. This problem has not been discussed enough," warns Naples University's Pagano.
This cuts to the heart of the difficulty facing eurozone policymakers. Ultimately, reducing banks' risky exposure to sovereigns must involve diversifying the investor base for sovereign debt away from the banks altogether.
"If you look at Italy, between the onset of the euro crisis in 2010 and the height of the crisis in 2012, a lot of external investors in Italian bonds left the market. They came back as the situation stabilised during 2013–14, particularly as quantitative easing became likely," says Reza Moghadam, vice-chairman for sovereigns and official institutions at Morgan Stanley in London, who headed the IMF's Europe department during the 2012 Greek sovereign debt restructuring.
"In fact, we have seen a widening of the investor base. We have now had both reduced volatility in the market, and a decline in yields and higher prices. In terms of the origination of sovereign bonds, since the middle of 2014 the market has been extremely supportive. For instance, Belgium was able to issue 50-year paper in April 2016, its largest ever syndication, which was vastly oversubscribed across a very wide investor base," he adds.
But Patel at UBS believes this more diverse investor base may not be particularly resilient. He says central banks and asset managers, especially in Northern Europe, reacted during the 2010–12 crisis by removing peripheral eurozone paper from their investment benchmarks. Although asset managers have returned to peripheral European government bonds, he says some have not yet restored these assets to their benchmarks. Consequently, their renewed investments in the periphery are off-benchmark bets, unlike the pre-crisis period.
"Were anything to genuinely upset that – a big market move or a big increase in volatility – then those peripheral government bond positions would be that much more mobile as a result of them sitting outside benchmarks rather than within them," Patel warns.
There may be less disruptive ways of encouraging diversification than via prudential rules. Santiago Fernández de Lis, chief regulatory economist at BBVA in Madrid – recently appointed chairman of the EBA's banking stakeholder group – thinks the ECB could consider introducing sovereign diversification requirements in the collateral framework for its monetary policies.
"Probably, the result would be similar, reducing the excessive home bias, but with a much less painful process," he says.
Private collateral flows could also face concentration limits. The EU non-cleared derivatives margin rules initially included diversification requirements on all collateral to be posted. But in June 2015, the European supervisory authorities (ESAs) decided to exempt sovereign collateral for non-systemic banks posting less than €1 billion in total margin. These are precisely the smaller banks where Fitch found home sovereign risk concentration to be highest.
"As many smaller market participants tend to have substantial investments in local sovereign securities and a diversification may increase, instead of reducing, their risk profile, the ESAs are of the opinion that concentration limits for this particular asset class should be required only for systemically important entities," the ESAs concluded.
Effectively, the ESAs hit the same problem outlined by Pagano of Naples University for sovereign concentration limits more generally – they increase contagion risk from peripheral to core eurozone banking systems.
A silver bullet?
If there is a silver bullet, it would seem to involve allowing banks to diversify their sovereign exposure without increasing contagion risks. Pagano, together with other academics and ESRB policy expert Sam Langfield, has co-authored an ESRB working paper that would revive an idea first proposed at the height of the eurozone crisis in 2011. This is to form a securitisation pool of sovereign bonds from all monetary union members, in proportion to their share of total gross domestic product, and distribute it in two tranches. The paper is due for publication in October.
"We show with some simulations that if you set the attachment point for the senior tranche to at least 30% of the total size of the asset pool, you can achieve a lower credit risk than German bunds," says Pagano.
This senior tranche, which the authors dub European Safe Bonds (ESBies), would therefore allow banks to diversify their holdings of sovereign debt without increasing the contagion risk. Moreover, it could permit diversification without the transitional problems of banks having to sell massive amounts of domestic sovereign debt.
The ESRB paper will suggest the ECB could jumpstart ESBies by participating in the initial sovereign bond/ESBies swap process and holding ESBies as monetary policy collateral, with a lower haircut than on individual sovereign bonds. Prudential regulators should also waive risk weights or concentration limits for ESBies while imposing them on undiversified domestic debt holdings, Pagano suggests.
"You would be guiding their transition to diversification with this pre-cooked safe portfolio," he says.
This still leaves the junior tranche to find a market. Pagano believes investors willing to take on extra risk for extra yield would be the natural buyers, including mutual funds, hedge funds, sovereign wealth funds and high-net-worth individuals. Pension funds and insurance companies might also be suitable, although the risk of bail-out feedback loops would need to be considered first.
"It is essential to avoid institutions that are highly leveraged, such as banks, or that have short-term liabilities and are assisted by a public guarantee, holding risky sovereign debt," says Pagano. He argues that breaking the sovereign-bank doom loop would make all sovereign debt safer – including the junior tranche of the securitisations.
The proposal will naturally prompt fresh questions, however. Pagano and his co-authors believe the securitisation would not require joint liability among participating sovereigns. By contrast, bankers interviewed for this article question whether ESBies would find a market without some form of multilateral guarantee similar to that enjoyed by bonds issued by the European Stability Mechanism.
Bankers in the eurozone say regulators have so far overlooked one important aspect of any new prudential rules on sovereign risk: the extraterritorial effects on non-eurozone activities. Spanish banks have large, often self-funding subsidiaries in Latin American markets such as Mexico, Brazil and Colombia.
"They hold domestic debt denominated in domestic currency, which is stripped by the local supervisor as zero risk weight. If you adopt a different policy in Europe, it would affect our consolidated balance sheet here in the eurozone. In this way, we would introduce risk weights in our holdings of domestic debt in domestic currency in these domestic subsidiaries that are funded with local deposits, that are supervised by the local supervisors, that are protected by the local deposit guarantee scheme," says Santiago Fernández de Lis, chief regulatory economist at BBVA in Madrid.
Austrian and Italian bank subsidiaries also have dominant market shares in central and Eastern Europe. Research for the European Parliament suggests Poland and Hungary are the two countries where banks have the highest home sovereign exposure in the EU (see figure 1).
"One of the key questions would be about foreign exchange risk. Clearly, it is much easier to diversify sovereign exposures within a currency union like the eurozone than outside. Therefore, outside, the choices banks would have around which assets they substitute would be different," says Alan Adkins, group credit officer for financial institutions at Fitch Ratings.
In particular, banks in non-euro countries would have to find alternative, high-quality liquid assets to meet their liquidity coverage ratio buffer. In more developed markets, such as the UK, this could include securities such as covered bonds, but in less developed central and eastern European markets these alternatives would not be available.
A further legal complication is the use of the central bank as the official issuing entity for sovereign bonds. Within the EU, this only applies to Denmark, but the practice is more widespread outside the EU.
"What happens in a country where they raise the central bank reserve requirements? Would you increase the capital requirements at the same time? You are forced to have these positions vis-à-vis the central bank, but if you don't include them [in prudential requirements], you may open regulatory arbitrage between the treasury and the central bank," says BBVA's Fernández de Lis.