Risk Management: Extreme risks
Tim Hodgson takes a qualitative approach to ranking extreme risks and assessing their interconnectedness
The events of the last two years have demonstrated that risk management cannot afford to stop at the 95th percentile (VaR95). We need to find a way to include unlikely, but potentially high impact, events. Dividing these extreme risks as financial, economic or political, we start with the qualitative understanding - ‘What could cause this event?', ‘Is it plausible?' and ‘Assuming this event occurred, what would be the consequences?'
Financial extreme risks revolve around solvency. The interconnected nature of the modern financial system and high levels of leverage mean that insolvency for one institution can quickly become a systemic problem. The primary triggers are falling asset prices and incomes. Financial risks can be self-generated (falling asset prices) and transmitted to the real economy, or generated by a recession in the real economy which reduces incomes and transmits to the financial sector through loan defaults. With public sectors increasing their leverage (debt to GDP ratios) we believe the risk of excessive debt will persist for a number of years.
Economic extreme risks are less homogenous, ranging from a deflationary depression to hyperinflation and a return to a gold standard. The deflationary depression risk appears to have been reduced through policy action, but remains a potential hazard, in that it may not be possible for governments to counteract any future drop in demand. There has been an extended period of over-consumption (Western economies) meaning that businesses have built productive capacity to satisfy a level of demand unlikely to be reached for a number of years, as households increase their savings rate. The primary consequence of a depression is a sharp, and prolonged, increase in unemployment, with effects including a drop in consumption, restriction of credit, shrinking output, investment and numerous bankruptcies. This would be bad for asset prices and it is likely that there would be a flight to the safety of sovereign (nominal) bonds.
The other economic risks essentially assume that government actions are successful, but at a price. A currency crisis - the breaking of a fixed exchange rate or expectation of a significant self-fulfilling devaluation, perhaps due to budget policy mismanagement or a lack of central bank reserves - could be profited from by holding unhedged foreign assets, but is likely to be disruptive to domestic asset returns, at least in the short term: the crisis will have been triggered by some form of economic malaise, reflected in lower asset prices. A currency collapse severely reduces a country's purchasing power and hence wealth. To the extent that domestic borrowing has occurred in foreign currencies, the cost of servicing that debt will rise, and hence increase the risk of default. At a more micro level, the equities of exporters should perform better than those of importers.
Should stimulating the economy require government debt to grow to an unsustainable level, then we would have the prospect of sovereign default or hyperinflation. Either of these would be bad for asset returns, including sovereign debt. (It is probably best to assume that inflation-linked bonds would be defaulted on in hyperinflation.)
Hyperinflation wipes out the purchasing power of savings, provokes extreme consumption and hoarding of real assets, causes the monetary base to flee the country, and investment ceases. The risks associated with sovereign default include penalty costs applied by external creditors on future finance and banking crises, as domestic banks tend to be large holders of sovereign debt. It is possible, that gold would act as a hedge against these risks, particularly as they would increase the likelihood of a return to a gold standard. We do not envisage that in the near future because escalating money creation is currently working fairly effectively to prevent a depression. However, should these efforts result in rising inflation, then the return to a gold standard would have a higher probability in future.
Our third category of extreme risks mostly derive from politics, but includes climate change and killer pandemics. There is a high degree of uncertainty attached to any estimate of the likelihood of occurrence, and these risks will also tend to be much harder to monitor and predict - with the exception, perhaps, of protectionism. In most cases, these risks will be hard to hedge. For example, hedging the break-up of the euro may involve the use of credit default swaps (CDS) contracts, which introduce different risks. Adding food and water exposure to a portfolio may hedge climate change but might run the risk of confiscation in precisely the circumstances in which they became most valuable.
Rather than attempt to give each risk a probability (the quantitative route), we assign them one of three risk categories - ‘low', ‘very low' and ‘very, very low' (figure 1).
We also assign each risk to an impact category: high implies a significant impact on most asset and liability values, medium a material impact on some values, and low an impact on a few values where the significance could vary. A subjective scoring system, combining the impact and the risk together with the degree of uncertainty in assessing the risk, produces a ranking (figure 2).
We also show an ‘association' matrix (figure 3). We use the term ‘association' rather than correlation to communicate that this is a qualitative assessment of whether events are likely to occur together rather than a quantitative assessment of past data. This clearly indicates that we see the financial risks and several of the economic risks to be associated. A second cluster can be seen between the political and economic risks: for example, we assess that protectionism and depression have a high association. The primary causality is likely to run from depression to protectionism but we would also argue there is a weak causality in the other direction, as greater protectionism would weaken economic growth. Also clear from the matrix is that climate change and killer pandemic are truly independent risks. They may be loosely related themselves, but otherwise only intersect with political crisis (where the causality could conceivably work either way).
This is an edited version of ‘Extreme Risks', a paper published by Watson Wyatt in November 2009. Tim Hodgson is a senior investment consultant at Towers Watson