Is more regulation the answer?
The Global Financial Stability Report* of the International Monetary Fund reported in October 2018 that the global financial system has become much more stable. Are we not fooling ourselves and becoming overconfident again?
Financial markets appear to be stable through the classical lens of absolute knowledge and manageability. However, using the gaze of complexity theory, the picture is not so rosy. In theory, moving to a system of more centralised management of counterparty risk between financial parties will lead to more control and stability. In practise, however, this assertion should be questioned.
Take, for example, the idea of a central counterparty. Since the 2008 crisis, regulators worldwide have forced financial institutions that are involved with transactions to execute them through a central counterparty clearing house, a so-called “CCP”.
Centralisation is good thing – right?
The good intention is that this centralisation of collateral creates a safer environment for parties (such as banks and pension funds) because the bilateral counterparty risk is replaced by mutual collateral. Through the CCP the direct relationship between buyer and seller ends after the transaction. The CCP becomes the new counterpart of both the buyer and the seller. If either of them does not comply with its obligations, the CCP – and its members – will honour the agreement and bear the risk.
Since 2008 almost all derivatives worldwide (over 75% currently) have been centralised through CCPs. Unfortunately, due to regulatory entry barriers, only a few large CCPs remain that take on all transactions and all risks.
It was assumed that all CCPs are 100% robust. But unfortunately, the reality is that we understand the world far less than we think we do. In practise, the dynamics of complex networks – which the CCPs are – turn out to work differently from what we expect. A relatively small shock can have disastrous consequences for a complex network such as the CCP.
When it goes wrong, it goes really wrong
In October 2018 a large part of the capital of all members of the NASDAQ Clearing CCP was swept away by one incident: a fluctuation in the energy prices market between the Scandinavian countries and Germany. Although the risk models had estimated a maximum price change of €4, it was, in fact, €5.56. In isolation, this price movement was surprising, but, in the context of market spreads, this movement was plausible.
This movement firstly led to the bankruptcy of one player, while the rest of the members hardly experienced any direct problems. But this single player did not only lose its own capital but also 70% of the reserve buffer of the joint pool of 166 members of the CCP.
By far the most disturbing aspect of this story was that the management of this CCP platform had no idea that an individual member of the CCP platform could do so much damage within the network. That is what they are looking for now.
What can we learn?
What worries me is that the regulators think they can control everything by centralising everything, but we do not understand how a relatively small shock could affect a complex network like a CCP. On a local level, if we do not understand something and it goes wrong, the effects are usually limited to the two parties involved. But here the problem is centralised and can affect an entire group or even an entire economy. We, therefore, would be wise not to hinge the world on a few networks and believe that we can manage them safely.
Centralising derivative transaction information was a good idea after the debacle in 2007 around credit derivatives in particular. Centralising transactions and their counterparty risk, thus interconnecting everyone via only a few institutes, is a step too far that has been taken too fast. These higher systemic risks are combined with an increase in liquidity risks, driven partly by CCPs having a strong preference for cash as collateral rather than government bonds, which are much more readily available.
We still have to find out what really big shocks will mean for the participants connected to CCPs. Even though the majority of members have good risk management, they too can be dragged down by network effects. A network model developed by VU professor Svetlana Borovkova^ shows that smaller players, such as pension funds, are more at risk within a CCP platform than when there is a bilateral settlement, i.e. no CCP.
Fewer players means greater risk
In a decentralised world with bilateral connections, unexpected things happen. In a centralised world with CCPs, unexpected things also happen. However, centralisation means risks are connected leading to more systemic risk and fragility of the overall system.
More understanding of network dynamics in finance will help the world far more than the traditional probability models. It will teach us that we know less than we think and that centralisation only yields a more stable world in very specific situations.
We need to be less arrogant about how much we can control systems. For now, will the regulators learn anything from the unintended side-effects of stricter controls? I hold my breath.
^ Svetlana Borovkova is a professor in the School of Business and Economics at the Vrije Universiteit Amsterdam