What’s the State of the European Banking System?

Q: What were the core drivers for the financial crisis in Europe?

The first crisis came about through investment in risky U.S. mortgage debt, which led to short-term liquidity problems. European banks were significantly invested in special purpose vehicles connected to U.S. mortgages, so what might have been a local mortgage crisis in the U.S. spread. As European countries had to raise large amounts of debt to bail out their banks, suddenly people realized sovereigns might become insolvent, too. Then they started to price risks.

In the decade leading to the crisis, you saw almost the same interest rates on sovereign debt across Europe. Greece, Spain, Germany, the UK, and France were essentially thought of as the same, and they were thought of as risk free.

Banks have to hold capital, which reflects the risks of the assets on the balance sheets. This is true for corporate debt, for all other types of securities, but when it came to sovereign debt, European regulation makes an exception. They said as long as you invest in European sovereign debt, which is originated in European currency, you do not have to hold any capital. It has, as we say it, a zero “risk-weight.”

A second exemption was made with respect to concentration limits. A bank lending to a corporate client can have a maximum exposure of 25% of its Tier 1 regulatory capital. For sovereign debt, this was lifted.

As the crisis started, the combination of pricing sovereign risk in the markets and completely ignoring sovereign risk on the regulatory side meant banks, particularly the large ones, were completely over-leveraged. It led not only to risk for individual banks but also to systemic risk in the sense that everybody was using the same theoretically risk-free type of asset. The consequences were amplified as banks tried to de-leverage, which increased downward pressure on asset prices that eventually spilled over into other asset classes.

This was particularly problematic as banks tend to be financed short-term. As investors were not willing to refinance banks with large holdings of sovereign debt, particularly peripheral debt, funding markets froze and the European Central Bank eventually stepped in with three-year Long Term Refinancing Operations.

This regulatory treatment of sovereign debt hasn't been changed yet.

Q: Why did the sovereign debt loophole exist in the first place and why does it still exist?

One of the biggest challenges for Europe compared to the U.S. is that we have 18 countries in the Eurozone and 28 in the European Union overall. They have different national accounting standards. The different national regulators have different rules and different incentives. All of these differences must somehow be accounted for in the common regulatory system that the European Central Bank is currently trying to create and oversee. And this of course involves political choices.

I think the sovereign debt regulation was an issue of helping countries refinance. It was a political choice. It’s also a political choice not to do anything to change it right now. There are people in the ECB who would like to address it, but they need empirical evidence.

What we need to think about from an academic perspective is how this can be changed in an efficient way. You cannot just change the rules from one day to the other. The banks simply don’t have enough capital. Moreover, how should bank funding be replaced?

A change to reflect sovereign debt as having something other than a zero risk would be an important issue for banks in the peripheral countries in the EU because they hold a lot of their countries’ risky debt. They are basically the only ones holding bonds of these countries at the moment.

For German banks, pricing risk wouldn't be much of an issue, but if we changed the concentration limits, it would create substantial problems. Banks have multiples of their own equity invested in not only sovereign debt but also municipal debt from within Germany, so this would also hurt not only large banks but also smaller banks such as savings banks.

We have to think about ways this can be phased out; perhaps there might be some grandfathering rules and new requirements apply only to newly issued debt, for example.

Q: Has this crisis threatened the EU?

Everybody's talking about the UK possibly exiting the European Union. But I think there has been the political ambition not to let anything happen to this goal of a united Europe. And that is probably not going to change.

The question now is, what are European governments willing to do to avoid the situation that we’ve seen in Japan for years? If Europe were to have a long period of no growth, that would put much more stress on this European project.

Also, how much is Germany willing to do in terms of investments and maybe paying something for the peripheral countries, which have more problems because they cannot regulate via their own currency anymore? If the core countries were willing to do more, it would be beneficial for the Eurozone as a whole.

Q: Could you talk me through the most recent set of stress tests?

Overall, it’s a huge improvement on 2010 and 2011. The fact that Dexia [a Franco-Belgium bank that was ranked among the safest banks in the 2011 stress tests and then went bankrupt a few months later] failed was a complete disaster and put the process into real doubt. This time the political stakes were higher for the ECB. They needed to show that they are a good regulator and not have anybody fail because of legacy assets that should have been detected.

That is why they started not only with the stress test, but a comprehensive asset quality review [AQR] that was supposed to look at non-performing assets within a unified framework.

The AQR collected information on about 50% of the overall risk-related assets of all banks. This hasn’t been done before, so on the one hand it’s very good they’re doing it. But on the other, it was very un-transparent; we know something happened, but no one exactly knows what the parameters are, what the assets are.

They made choices in terms of what assets to look at, and there might be risks in asset classes that weren’t examined. There's always a risk that the next crisis will come because of the stuff we overlooked.

Additionally, how do you put a valuation on the assets that were examined? This is always a very subjective part. There's no objective measure to get an idea how much a bank’s shipping loans are worth, what a real estate portfolio is worth. The valuation issue is very important and very subjective.

Q: Who made those judgments?

There is again an issue of 18 different countries being involved. The ECB has its own staff and consultancy companies, but as long as they are dependent to some extent on information coming from banks and national regulators, there's a huge issue with the incentive to reveal bad assets. Why would they?

Ideally, you would have done the AQR, and then used the same assets with the stress test. But, because of time constraints, they ran in parallel. The AQR was being done by the ECB in coordination with a couple of other teams. The ECB and the ESRB [European Systemic Risk Board] designed the stress test scenarios, handed them to the banks and said, "Here, run the test and give us the results."

The stress test results were adjusted to the bank’s AQR results, but the stress test didn’t actually use the same assets. The adjustment was done later in a third step called the join up.

That meant a lot of discretion for the ECB to get the outcome they wanted. And it’s not really clear ex-ante that they actually would like to have shortfalls resurface when there isn’t a clear solution for how those would be refinanced.

I have some doubt that a lot of banks would be able to refinance their debt in private markets, especially if there are really large shortfalls. Those would likely put governments on the hook.

Even as we are making progress in terms of recovery and resolution directives, which has been or is currently being put into place in the different nations, fundamentally we do not actually have a really well-defined backstop, which means no one wants to see shortfalls.

Q: Describe your research on capital requirements.

With Viral Acharya of New York University, I tried to come up with some benchmarking. Since we didn’t have the information about the banks’ balance sheets and we don’t want to speculate about scenarios that would lead to the next shock, since we're always wrong about that, we used different methods.

During the 2008-2009 financial crisis, how much capital was held by the banks that did not have problems? The Bank of England estimated for small banks it was 4%, and for large banks 7% book equity to total assets that got them through the crisis without any kind of government bailout. So we estimated how much capital banks need to come up to those thresholds now. We found that the European banking system needs an additional €400-500 billion in capital, while the ECB estimated a shortfall of €25 billion for all banks. An advantage of this method is that we actually avoid estimating losses during a crisis.

We also used “SRISK” as alternative measure of capital shortfall as calculated by NYU Stern School of Business Volatility Lab. This concept and the stress test conduced by the ECB are conceptually similar as they estimate losses in a stress scenario and determine the capital shortfall between a prudential capital requirement and the remaining equity after losses. The stress scenario is a systemic financial crisis with a global stock market decline of 40% over a period of six months and assuming a prudential capital ratio of 5.5%. This scenario and the resulting SRISK measure use market data and market equity, instead of book equity, in determining leverage. We estimated a capital shortfall of about €500 billion just for the European public banking sector, where the ECB estimated a €20 billion shortfall.

While our benchmark estimates had a negative correlation with capital shortfall, SRISK is positively correlated with total losses incurred by banks in the adverse scenario. This suggests that it is not the actual losses but rather different ways of specifying prudential capital requirements that must be driving the negative correlation between SRISK and the results of the ECB. To be precise, our analysis shows that it is the use of risk weights in the regulatory benchmark that explains the shortfall differential between our and the ECB’s assessment. Future stress tests should incorporate a robust approach that does not rely exclusively on risk-weighted assets but adopts multiple approaches including a simple leverage ratio.

Q: Why are banks able to run on these leverage ratios?

No corporation would run on a leverage ratio like a bank does. One reason it’s tolerated is that banks are so important for everybody that we know if there’s a problem we will rescue them. Everybody knows there will be a bailout. If there were no chance of a bailout, nobody would allow banks to work with these low capital ratios. Bond investors would demand much higher ratios.

I think banks need to have more capital, and more of the debt on the bank balance sheet should be outside the banking system. Banks are the largest holders of other banks’ bonds. In Germany, about 80% of all bank bonds are held by other German banks. As long as this is the case, I doubt the banks will be able to access the capital they need in a crisis which, of course, means that taxpayer money is still at risk.

Q: How much is systemic risk still an issue?

What the ECB was actually trying to do with the stress test was to look at the assets and determine whether each bank is individually well capitalized. They haven’t looked at what happens to the system if one bank needs to deleverage. The interconnectedness between banks, between banks and markets, between banks and the real economy, and the feedback effects back to the financial sector are complex and is not covered by this test. It’s something which needs to be changed over the next couple of years.

Q: Why isn't evaluation of systemic risk built in?

Consider whether running the current stress test in 2006 would have mitigated the fallout from the 2008-2009 financial crisis. Probably not, because nobody would have realized that, for example, AIG was a substantial source of risk for all European banks. It’s difficult to really look at systemic risks. Ideally, you consider not only banks but insurance companies and the shadow banking sector.

Maybe it’s not possible to put this into the sort of quantitative models the ECB has right now. But a possibility could be to do something like what the Federal Reserve does which is to do a qualitative test where you ask questions like, “What happens to your bank if AIG defaults?” Then you get a much better idea about how interconnected the banks actually are and maybe come up with a completely different capital requirement.

It’s not only important to know the counterparty of the bank—regulators have some of this information already though it's not used to the full extent it should be. It's also important to know the counterparty’s counterparties. It’s important to know how broad the network is. If you pull the plug here, what happens to the rest of the system? There are lots of techniques available in economics and social sciences to measure networks which could be applied by national banks which have the necessary depths of data.

As a result of the regulatory data that is available now, the ECB is in a much better position to look at this than anybody before. They can compare business models of large banks such as Deutsche Bank, Société Générale, and Crédit Agricole. It is probably the first time that a regulator has better information than the banks themselves.

They have an information advantage over the national regulators, even the banks, because they can see everybody. It puts them in a unique position to come up with a network topology. The next couple of years will be interesting times to work at the ECB as they get this data together.

Q: How has your critique been received?

Well, I think. I presented this at the European Parliament, and they were actually very open. We had a very long discussion on the systemic risk component.

The ECB might have been a bit more skeptical, but I think it was well received. The stress test is going to be done on a regular basis and the next tests will show whether and to what extent these ideas have been implemented.

Q: What needs to happen going forward?

I think a possible next step in terms of capital regulation might be to think about the treatment of sovereign debt. I think this should be highest on the political agenda.

Another thing would be to make risk weights less static. There are questions about how this can be done. We need to think about how we can measure risks on a regular basis in a good way, avoiding procyclical effects.

Q: How about bigger economic issues?

We need to think about both the supply and demand sides. One reason for the AQR and stress test was to get a handle on the supply side. If banks are still completely over-leveraged with bad assets that they cannot get rid of and they cannot write down because they don’t have the capital, they can’t lend to healthy parts of the economy. They would rather extend debt to the old risky counterparty in hopes that counterparty doesn’t default.

On the demand side, there's risk in the Russia-Ukraine conflict or ISIS. That might prevent firms from investing in the short term. But for Germany, I think there’s too much focus on a balanced budget without looking at the health of the economy. We need to spend some money to revive the economy on a short-term basis and then implement the structural reforms that could actually support a longer-term and more sustainable growth path over the next couple of years.

Germany is in a very good position to spend some money and it could have a huge impact on other countries, so perhaps other heads of countries would start doing it as well.

Q: How much will is there to start spending?

Not too much, apparently.

Interview conducted and edited by Ted O’Callahan.