resources Article

Simulation to Explore
Contagion Risk

Contagion risk is the risk that a shock to one financial institution spills over to others. In this way, small shocks can have significant effects. Contagion is one of the key dynamics that gives rise to systemic risk in a complex adaptive system. The complex web of interconnections between the constituent parts of the financial system can act as pathways along which instability spreads and amplifies.

The web of interactions between financial institutions is vastly complex.

This understanding of contagion risk is not merely theoretical. During the financial crisis, fears about possible contagion led the US Government to extend an $85billion bail-out to AIG. It wasn’t so much the fact that AIG was “Too-big-to-fail”, but rather that it was “Too-interconnected-to-fail”. AIG had quietly gone about building a significant CDS (Credit Default Swap) business. These complicated derivatives were used by other financial market players, like banks, to insure against loans going bad.

Practically every financial institution on the planet was using AIG’s CDS to protect itself. The failure of AIG could have led to chaos as previously insured loans suddenly became risky again. The chain of events that removing this key player would trigger were unknown. By one estimate, the failure of AIG would have cost its swap counterparties $180 billion — potentially triggering further failures, and further rounds of contagion.

Liquidity Risk & Contagion

The 2008 financial crisis was characterized by a global liquidity squeeze.

Banks provide a key function in society: they perform maturity transformation. That is, they lend money long-term (for example by providing mortgages), yet they stand ready to provide cash on demand to their depositors. In doing so, they generate liquidity risk — the risk that they will be unable to fulfill their obligation to provide cash on demand because their assets are locked up for the long term. Such runs can trigger insolvency by forcing banks to “fire-sell” liquid loans at below-market value.

In the traditional banking model, liquidity risk stems from the possibility of bank runs — episodes in which a large number of depositors try to withdraw their money simultaneously. These can be triggered by concerns about a bank’s health, or a self-fulfilling prophecy as a product of a simple coordination failure: households think that others will try to withdraw cash and don’t want to be left high and dry. This type of depositor run is rare in modern banking thanks in large part to deposit insurance which removes the incentive for households to panic and withdraw funds en masse.

Instead, the modern bank run — and hence the modern source of liquidity risk — is a result of exposures to a plethora of lending and interbank financial arrangements. Banks finance their balance sheets in increasingly complex ways. The traditional retail bank, funded with retail deposits and equity capital, has been consigned to the textbook. In the real world, banks fund themselves with a range of liabilities, which include uninsured wholesale deposits, repurchase agreements, and other short-term unsecured debt instruments.

The case of Northern Rock is the modern bank run example that will be taught to the next generation of economists. Northern Rock was a relatively small regional lender which underwent rapid growth in the build up to 2007. It was heavily reliant on short-term funds which dried-up suddenly following a re-evaluation of credit constraints by its counterparties. Do not be fooled: the queues outside branches were merely a symptom, not the cause of the failure of Northern Rock.

By becoming increasingly reliant on other financial institutions for its funding, Northern Rock had exposed itself to shocks in other parts of the financial system. What started as quiet whispers about US sub-prime markets, brought down a regional UK lender. This was a powerful illustration of contagion at work. A seemingly small and distant trigger was transmitted across the globe and amplified as it traversed networks of financial contracts.

Overlapping supply chains

Another example of contagion risk has emerged in recent weeks here in the UK. The construction company Carillion has collapsed, leaving thousands of suppliers facing the fall-out of unfulfilled contracts. This could cause other firms to get into difficulty, again setting off waves of financial distress. Lenders who had lent, not only to Carillion, but also to its suppliers, suddenly realized they were going to be hit with missed repayments twice — first by Carrilion, and then by its suppliers.

Traditional correlation analysis may not pick this up. Without knowing that a firm was a supplier of Carillion, it can be hard to detect a relationship between the two. Without knowledge of the financial network, it’s not possible to uncover a causative contagion relationship.

This is just one more example of the maxim that modeling at higher levels of aggregation and considering only correlations, rather than causative relationships, can miss important dynamics.

The crisis of complexity

As balance sheets have got more complicated, the webs of interconnectivity have become increasingly tangled. In a globalized world it is no longer enough to consider risks to a bank as though it were isolated from events elsewhere in the financial system.

Recognizing that the financial system is a complex adaptive system allows us to account for the types of dynamic that were witnessed during the crisis. And putting financial networks at the heart of that thinking allows us to move beyond considering banks as isolated nodes, and start thinking of them as part of a larger system, the properties of which emerge from the interaction of its constituent parts.

In response to this, we need simulation models that explicitly capture low-level heterogeneity, and explicitly account for the interconnectivity between the constituents of the financial system. Agent-based modeling for risk management is increasingly showing itself to be the way forward. Combining agent based modeling to simulate network analysis allows us to capture causative relationships — be they funding providers pulling funds from banks, CDS contracts linking financial institutions to failing insurers, or suppliers losing out on contracts from failed construction companies — that have the potential to do significant harm.

Chloe Hibbert