European banks are the main source of a global systemic risk – claims Professor Plutarchos Sakellaris from the Athens University of Economics and Business.
He proposes a new method for determining the systemic importance of banks, taking into account the volatility of prices of their CDSs and not only the size of a given institution. The simplest method of determining the systemic importance of a bank is the size of its assets. We can also establish that risk-weighted assets are crucial for the bank’s position in the financial system. Another method draws attention to the size of the bank’s leverage ratio, showing how many times the leverage used by the bank exceeds its equity.
“Size is not the only thing that matters. The composition of assets and the government’s ability to support the bank are also important,” said Plutarchos Sakellaris during a seminar at Poland’s central bank, Narodowy Bank Polski in Warsaw.
In 2010, Christian Brownlees from Universitat Pompeu Fabra in Barcelona and Robert Engle from the New York University published the assumptions of the method known as SRISK. It takes into account more factors: the size of the bank, its leverage and the expected capital shortfall in adverse market conditions.
After the recent crisis, it was found that huge banks, which have been growing since the 1990s as a result of mergers and acquisitions, are of essential importance for the global financial system. The collapse of such a giant bank would have unimaginable consequences for the financial system, and the bailing out of such a giant would require unimaginable resources. Therefore, in 2013, the Financial Stability Board (FSB) and the Basel Committee on Banking Supervision (BCBS) developed a methodology and each year they identify 30 global systemically important banks (G-SIBs). Due to their systemic importance, these institutions must fulfill capital requirements increased by an additional buffer, of up to 3.5 per cent.
Plutarchos Sakellaris and Georgios Moratis selected the 150 largest banks in the world. Out of these banks, they identified those whose issues of debt securities are insured by credit default swaps and are traded on the markets. In 19 developed countries and 7 countries classified as the emerging economies there were 77 such banks. They then examined the daily spread volatility of 5-year CDSs in the period from January 2008 to June 2017. “The CDSs are a very good instrument showing increases in credit risk,” said Plutarchos Sakellaris.
Crises are spreading through the network of mutual connections
What do the CDS quotes show? An increase or decrease in the credit risk of the institution whose debt they insure. Let’s say bank A has an exposure to bank B. The credit risk of bank B is growing, which is reflected in its debt spreads and the cost of insuring that debt, i.e. the valuation (price quotations) of the CDSs. In this way, the risk of bank A is also growing. If bank B turned out to be insolvent, the exposure of bank A to bank B would be lost. As a consequence, the risks for bank A due to lost exposures to bank B spread to other banks, which in turn have some exposure to bank A. Each bank transfers risk to others and at the same time takes on risk from other institutions.
Mutual exposures of banks create the interconnectedness between them. These are connections that determine systemic risk, and they do so to a much greater extent than the size of the institution. The size of mutual connections is burdened with the risk of the other party. The CDS spreads serve as a measure of this risk. “Systemic importance is strongly linked with CDSs,” said Plutarchos Sakellaris.
Let us further assume that we are not dealing with a huge, global shock that occurred during the last global financial crisis. The shock occurs only due to the specific risk of one bank, i.e. idiosyncratic risk. But this shock could spread to the financial system. All that is required for a systemically important bank is to succumb to it.
“In a global system the connections between the banks result in the transfer of the risk across the network. In this way idiosyncratic shocks can also move to the systemic level,” said Plutarchos Sakellaris.
Systemic risk is therefore not a consequence of a bank’s size. If we were to consider the world’s three largest banks in terms of assets: Bank of China (USD2.613 trillion in assets), Mitsubishi UFJ (USD2.597 trillion ) and JP Morgan (USD2.490 trillion), it would turn out that in the global financial system they do not matter as much as would be indicated by the size of their assets. Banks that are much smaller, but more connected with others, are much more important in the financial system. And at the same time the latter are also perceived as more risky, as shown by the prices of CDSs for their debt. They also pose a greater threat to the entire system.
Based on the mutual connections and the bank’s risk determined by the spreads of CDSs on its debt, it is possible to measure risk transfers between the banks. Such risk transfers can be presented in a “contagion table”, which presents the results of a given institution’s external risk factors.
How does the BCBS measure systemic importance
Every year in November, the Financial Stability Board announces the list of the 30 largest global systemically important banks (G-SIBs). The list is based on the methodology of the Basel Committee on Banking Supervision. Banks that are included on the list must have their capital requirement increased by a buffer for global systemically important institutions, ranging from 1 per cent to – theoretically – 3.5 per cent. The buffers are introduced gradually and will become fully effective from 2019. Thus far, there has been no bank that would be required to have the highest buffer corresponding to the fifth bucket. JP Morgan is the only bank in the fourth bucket with a buffer of 2.5 per cent.
The BCBS methodology is also based on the measurement of the institution’s size, although there are many measurement criteria. The inclusion on the list is not determined solely by the size of assets, although it is one of the criteria. The scale of operations and the bank’s risk are described by 12 indicators which are in turn divided into 5 categories.
Each category is assigned the same weight of 20 per cent. The individual indicators that make up the categories also have weights assigned to them – 10 per cent or 6.67 per cent. The indicators include: the size of cross-jurisdictional claims and liabilities; assets and liabilities connected to the financial system; the portfolio of marketable securities; assets in management; illiquid level 3 assets; the volume of the implemented payment transactions and the share in the infrastructure, e.g. of the payment systems; the value of the underwritten transactions in debt and equity markets; the securities held for trading and available for sale; and exposures to derivative instruments.
Then, each of the indicators for a given bank is divided by the value corresponding to this indicator for all the others from the sample. These values are announced annually. For example, the value of JP Morgan’s cross-jurisdictional claims is divided by the value of the cross-jurisdictional claims of all the banks considered. The result of this operation is converted into scoring points, which in turn qualify the institution for the appropriate “bucket”, for which an adequate buffer is provided.
Systemic risk follows the CDSs
Credit default swaps are instruments which are particularly sensitive to risk. From 1994, when they have been introduced to the markets by JP Morgan, until 2008, they were a great business for the insurers. Why? The economy was booming and the bankruptcies of securities issues were rare. This was the case until the first decade of the current century, when synthetic CDOs (collateralized debt obligations) securitizing mortgages in the United States began to be insured.
According to data from the International Swaps and Derivatives Association (ISDA), at the end of 2007 the value of outstanding CDSs on the markets reached USD62.2bn. Because the trade was not transparent, many analysts and economists believe that this instrument has led to the spread of systemic risk through a network of financial institutions.
There is one more problem associated with CDSs. It is based on the fact that the insurance of the underlying instrument does not have to be connected to any participation in the trade activity regarding this instrument. We can buy some issue of instruments and not buy the CDSs that insure it. We can buy or sell CDSs themselves, and not have in our portfolio any security which they insure.
One of the biggest issuers of debt insurance before the crisis was the London branch of the financial products division of the American company AIG. As the issuer of securities – in contrast to the banks or hedge funds that could take positions on both sides of the transaction, and therefore their net position resulted from the mutual offsetting of both sides – AIG only had liabilities to the insurances of the CDOs. When successive series of these securities turned out to be insolvent, the investors demanded the payment of insurances from AIG.
AIG – as it later turned out – had liabilities in respect of bonds worth USD440bn, including almost USD60bn in liabilities in respect of structured securities for sub-prime mortgage loans. As the issuer of the CDSs for these securities, AIG took on a substantial part of the systemic risk that materialized at that time. This led to the bankruptcy of AIG, which was bailed out with public funds.
Europe is extremely important
What are the conclusions from a methodology based on the study of links between banks and the transfer of risk expressed by the CDS spreads? The results are surprising. The most systemically important bank is not one from the G-SIB list, but the Italian Intesa Sanpaolo, a bank with assets of “only” USD766bn. Intesa Sanpaolo is responsible for 2.25 per cent of the overall global risk.
Next, we deal with the first G-SIB – the Spanish Santander Bank (the owner of the Polish BZ WBK). On the G-SIB list, it is only included in the first bucket. Santander is responsible for 2.21 per cent of the global risk. The next systematically important bank is the Spanish BBVA, which was included among G-SIBs for some time in the previous years. The next two banks are BNP Paribas and Unicredit. They both currently belong to G-SIBs.
All the places, from the 1st to the 20th, are taken by banks from Europe. The first American institution on the list, Morgan Stanley, was only ranked 21st. Although the G-SIBs account for 48 per cent of banks in the first quartile of the “global risk table”, they also account for as much as 14 per cent of the last quartile. Banca Monte dei Paschi di Siena, which was recently bailed out by the Italian government under the “precautionary recapitalization” procedure and is the 68th institution in terms of assets, was ranked 16th, ahead of all the American giants.
“The systemic importance of smaller European banks, such as Banca Monte dei Paschi di Siena, is growing,” said Plutarchos Sakellaris.
While in the first quartile, i.e. among the top twenty most systemically important banks, there are only European institutions, the American giants only appear in the second quartile, and there are eight of them, which means they represent 40 per cent of that quartile. In the second quartile, there are only five institutions from Asia, i.e. 25 per cent.
By applying this methodology further, we can also calculate the systemic risk generated by European banks. They are responsible for 69 per cent of the risk that other European institutions “take on” from the network of global connections. But they are also responsible for 58 per cent of the external risk of American banks and 41 per cent of risks of institutions in Asia.
“Most of the shocks ‘received’ by any region of the world come from Europe. And vice versa – the majority of shocks that may occur in another region come from Europe. Europe is extremely important,” said Plutarchos Sakellaris.
What else can CDSs show
Soon, the valuation of credit risk reflected in CDSs may become even more interesting due to the issuance of new debt instruments that are already hitting the market. This will be debt classified as MREL and TLAC (for G-SIBs). The Spanish, French and Italian banks have already carried out the first such issues last year.
The idea behind the MREL and the TLAC is that in the case of insolvency of the institution and the supervisory body’s decision to subject it to resolution, liabilities included in these categories (according to European law this debt is supposed to constitute a new category of non-preferred senior debt) are supposed to serve the recapitalization of the bank. To greatly simplify the matter, the previously held equity will be used to cover the losses, and the debt classified as MREL or TLAC will be converted into new equity so that the bank subjected to resolution is able to operate on the market.
Credit default swaps will also be issued for the debt classified as MREL and TLAC. At the end 2017, the ISDA introduced appropriate changes in the regulations. They will probably be more expensive, but also more sensitive to the risk of the issuer of the securities which they insure.
Gavan Nolan, a director at IHS Markit, writes that the behavior of the debt included in the MREL of European banks (such as BNP Paribas or Barclays) has so far shown that spreads between this debt and senior debt are significantly expanding during periods of risk aversion. On this basis, it can be concluded that the CDS spreads in such periods will be much wider. “We can expect that the valuation of CDSs will be subject to similar fluctuations (as in the case of the underlying securities) and could be 70-100 per cent wider from CDSs for the (currently listed) senior debt,” wrote Nolan.
CDSs for debt classified as MREL/TLAC should be an instrument which is even more sensitive to credit risk than the insurance of senior debt. The question arises, however, of whether they will only serve as a tool for risk measurement or as a new risk transfer instrument? And if so, in what direction?
The BCBS realizes that TLAC instruments could create a new channel of “contagion” in the banking system. Because of that, it introduced a threshold limiting the bank’s holdings of this type of securities issued by another credit institution to 5 per cent of the investing bank’s equity. But no limits have so far been introduced for CDSs on non-preferred senior debt.
One additional interesting question arises in connection with the new debt instruments. The CDS issuer assumes responsibility in the event of bankruptcy of the issue of debt for which it has issued the insurance. If the same applies to MREL/TLAC debt, then in the event a bank’s resolution takes place, the holder of its non-preferred senior debt may apply to the insurer to fulfill its obligation. In this situation, the shareholder in the new equity would but the issuer of the CDSs and not the creditor. And this means that risk will start to migrate into completely new areas.