[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Economics: The road to ruin

Over the past 18 months, the banking sector has reported unprecedented losses arising from reductions in the value of assets on their books. Significant write-downs were made to super senior collateralised debt obligation (CDO) positions with underlying collateral including US sub-prime residential mortgage-backed securities (RMBSs). The impairment was attributable to higher than expected delinquency rates and a fall in house prices underlying sub-prime mortgages, causing credit deterioration of sub-prime RMBSs.

Reflexivity and failed risk management
Although risk managers were aware of these balance sheet positions, the risk management framework failed for a number of reasons. Firstly, reliance on the external AAA rating, which is the highest credit rating category available from rating agencies, to support underwriting decisions. The expectation of a low probability of default and high recovery rate associated with AAA ratings created a false sense of security.

Secondly, failures in risk identification and measurement led to an underestimation of risk. The high initial rating gave the perception of a high return on economic capital. The Bank for International Settlements (BIS) recently issued a consultation paper proposing enhancements to the Basel II framework to strengthen risk capture. Improvements relate to risks overlooked in previous capital calculations — for example, credit migration, spread widening, liquidity horizons and the higher systematic risk of CDOs of asset-backed securities (ABSs) positions. The banking industry has been criticised for highly leveraged balance sheets. The underestimation of risk and capital requirements was a major factor contributing to this higher leverage.

Thirdly, inadequate risk mitigation caused problems. The availability of monoline credit protection on these super senior tranches reinforced perceptions that this risk was ’insurable’. As a result, some banks used the ‘negative basis’ trade, which involved buying the super senior tranche and then buying credit protection from a monoline with the aim of targeting a hedged ‘risk-free’ positive carry to the bank. However, the problems relating to counterparty capacity and the high correlation of counterparty risk to the underlying collateral were overlooked — for instance, ‘wrong way exposure’ whereby as the underlying collateral deteriorates, so does the monoline’s rating. In practice, the risks associated with these positions were non-hedgeable risks. Had this been recognized and sufficient provision made for the cost of additional capital, the return on economic capital for the underwriting of these credit positions would fall, to act as a compensating control on the volumes being underwritten.

Fourthly, during a short period from 2005 to mid-2007, market price data and volumes of new issuance supported the appetite for these instruments and expectations of performance. The herd mentality of investors caused by a ‘reflexivity’ feedback loop led to outcomes that were self-fulfilling, but inevitably self-defeating, in a boom/bust sequence. Market participants were using inadequate historical data to set their initial risk perceptions.

Finally, a failure to appreciate the interaction between risks did not help. Investors in short-term commercial paper that financed off-balance sheet conduits and structured investment vehicles exposed to RMBSs reallocated their short-term investment to short-dated government debt in a flight to quality. Under this liquidity stress, some banks decided to support the vehicles they created by consolidation, which resulted in an increase to their capital and liquidity requirements. The subsequent write-downs on credit investments also led to mistrust between banking institutions which caused significant problems in the inter-bank lending markets in 2007 and 2008. As a result, the banking system suffered from a combination of withdrawal of liquidity and falling solvency due to credit write-downs.

Knightian uncertainty
The lack of correspondence between original expectations and eventual outcomes reflects a failure to properly appreciate the uncertainties that were contained in underlying sub-prime risk. The great economist Frank H Knight distinguished between risk and uncertainty in his seminal work, Risk, Uncertainty and Profit (1921). Knight examined the concept of risk and made an important distinction between “risk as randomness with knowable probabilities” and “uncertainty as randomness with unknowable probabilities”.

Knight’s concept of uncertainty offers a possible insight into how this current credit crisis developed. The sub-prime mortgage market was more heterogeneous and had either shorter or worse credit histories, if any, than its prime counterparts. In order to generate credit rating estimates or risk assessments on mortgage-backed securities (MBSs) it is necessary to form a view on the probability of default, loss given default and the prepayment rate, as well as the correlation and timing of default of the underlying sub-prime portfolios. The fact that there was very little relevant or even reliable comparable data or history to support these estimates did not deter the origination of MBSs with collateral pools containing this sub-prime uncertainty.

Post-issuance data has shown a far higher realisation of delinquencies on sub-prime mortgage portfolios than initially expected. The new loss projections are extreme enough to downgrade some previously AAA-rated senior US sub-prime RMBSs to below investment grade or ‘junk’ status. Knightian uncertainty cannot be easily priced by markets because it relates to fat-tail distributions and extreme events that are difficult to predict and measure. The credit markets are currently very illiquid and market participants face the difficulty of determining the valuation for some troubled financial instruments.

Valuation of troubled assets
In March 2009, the International Accounting Standards Board (IASB) updated disclosure requirements on financial instruments under IFRS 7. This standard applies to all companies that hold financial instruments. It establishes a three-level hierarchy for making fair value measurements:

>> Level one
Quoted unadjusted prices in active markets
>> Level two
Inputs, other than quoted prices included within level one, are inputs that are observable either directly as prices or indirectly derived from prices
>> Level three
Unobservable inputs are inputs that are not based on observable market data. In the current illiquid market conditions, level three is most appropriate for many troubled assets.

Other IASB guidance states that valuations based on management’s internal assumptions that do not take into account factors such as credit and liquidity risks would not meet the objectives of fair value measurement. In practice, the valuation methods that may be considered in these illiquid markets include (i) cash flow simulation and/or (ii) a market-value approach that accounts for the intrinsic and extrinsic value of tranches based on collateral value. Valuation methods should use relevant observable market information where possible, and the market participants should agree observable and unobservable inputs. The price discovery process should not be ignored or put off as it may contribute to continued uncertainty and further deterioration of pricing and risk factor assumptions.

Reflexivity appears to have fuelled a bubble in the financial markets. Further, the failure to focus on the limitations of data, pricing and risk models for US subprime RMBSs and also CDOs of ABSs led to uncertainty being misinterpreted as a measurable risk. The resulting poor business decisions have caused unprecedented losses in the financial system and difficulties in price discovery.

Francis Richard Pereira is an actuary (FIA) and chartered accountant (ACA). The views expressed in this article are his own personal opinions