ERM Lessons from the Credit Crisis
By Jonathan Bilbul, U.K. Correspondent
Which car can travel faster around a race track, one with brakes or one without? A car will face many obstacles, let alone many bends in the track before it reaches the finish line. The ability to brake allows the race car driver to slow down to meet these challenges and to accelerate only when there is the most gain to be had. Similarly, companies want to be resilient in the face of risk and also to be able to exploit it should opportunities for gain arise. Especially in the financial community, an enterprise risk management system that is quick and responsive to change is central to ensuring success.
Solvency II, the new regulatory regime to be implemented across the European Union in 2012, recognizes the ability of insurance companies to use risk management systems to their benefit. The Use Test of internal models ensures that risk measurement and key decision making processes throughout the organization are aligned. Much more than a regulatory requirement, having a handle on a company’s risks can be a recipe for success if the information provided is relevant and timely.
Insurers can learn much from banks, our sister companies in the financial sector, and the hardship they have experienced in the recent financial turmoil. Banks also use internal models to measure risk associated with their activities and for management decisions.
However, the banks’ models came into question when daily trading losses in the third quarter of 2007 exceeded the possible losses indicated by the Value at Risk (VaR) measure at a 99% confidence level on several occasions. While two or three such occurrences can be expected in a typical year, some investment banks reported as many as 16 exceptions. If the internal models do not accurately represent the risk inherent in positions taken, then the models cannot be effectively used to guide management decision.
Some examples will help illustrate enterprise risk management failures as demonstrated by the current credit crisis. First, two failed business strategies will be examined.
The credit crisis triggered a run on Northern Rock, the first time this had occurred on any British bank in 140 years. While the bank was always technically solvent, with asset values exceeding liabilities, it faced a liquidity problem. Northern Rock’s strategy placed far greater reliance on money markets to fund its mortgage lending than any other retail bank. When investors lost their appetite to finance any mortgage-related activities, whether subprime or not, the bank was no longer able to meet its required payments.
In September 2007, Northern Rock sought emergency funding assistance from the Bank of England, the lender of last resort. By last February, the government had provided £25 billion in loans and £30 billion in guarantees. The ultimate solution was to nationalize this troubled bank.
Bear Stearns was another victim of the fallout from the subprime mortgage crisis. Its business model was highly dependent on the U.S. fixed income market. While this investment bank flourished from 2001–2007 when interest rates were low and the housing market was booming, its luck ran out when demand for securities backed by subprime mortgages faded. Its capital cushion of $17 billion evaporated within a matter of three days from March 13 through March 15, 2008. This led JP Morgan Chase, backed by the Federal Reserve, to make a buy-out offer at $2 per share (later raised to $10) when its stock had traded as high as $171.51 a year earlier.
The failings of Northern Rock and Bear Stearns show how strategies are doomed to fail when expected future financial conditions fail to materialize. To put it another way, they both hit the crash barrier because they failed to apply the brakes at the right time. In contrast, Lehman Brothers, another large investment bank, had also been heavily dependent on the fixed income markets about ten years earlier. However, it took steps to diversify its revenue sources and has not suffered the same fate as Bear Stearns.
Further examples of the difficulties faced in the credit crisis are found by examining the stories of large monoline insurers, such as MBIA and Ambac. The financial strength of these companies was questioned since they underwrote insurance policies that promised to indemnify bondholders from issuer default, certain of which were for mortgage-backed securities. Their top notch, AAA credit rating was placed under review by the rating agencies and any downgrade might have had severe effects on the cost of borrowing in the larger economy.
These insurers guarantee, amongst other things, debt issued by governments to build hospitals, roads, and schools. Any downgrade of the insurers would imply increased borrowing costs for bond issuers who acquire the same credit rating as their insurer. Additional capital had to be raised by both companies by issuing stocks and bonds and eliminating dividend payments in order to reinforce their financial strength and maintain their credit ratings.
These are only a few examples of the difficulties faced during the credit crisis. In the first two cases, the companies ceased to exist as independent firms, while in the third, certain monoline insurers were forced to seek additional capital from various sources. Why have internal models not been more robust at estimating the risk inherent in companies’ activities? The recent prior years used to calibrate these models were of benign market conditions. Models used by banks that incorporate results from one to five prior years of history would not reflect the volatility and extreme events of the second half of 2007.
There are two possible solutions: either build models that are more responsive to current conditions or calibrate models over a longer time horizon to incorporate a more realistic level of volatility. In either case, judgment is required to assess the appropriateness and completeness of the data used.
According to the Use Test under Solvency II, internal models should reflect the risk profile of the company and be based on current information. However, this is easier said than done. It requires a slick and quick process that allows for the latest data to be continually incorporated into the calibration model. The model should be compared, on a regular basis, to actual experience in order to validate its accuracy, but the user should also question whether deviations of actual from expected are sufficient to warrant a change in parameters or assumptions. For example, a property and casualty insurer that sees poor underwriting results emerge in the first quarter might cause it to question the realism of its business plan and the likelihood of the various possible outcomes. In response, the insurer should take proper action to mitigate unfavorable outcomes and their resulting effects on capital.
Another example relates to price inflation. With the prices of oil, gold, and wheat achieving new heights and unusual levels of volatility, inflation has become more difficult to predict. Current data will change estimates of future inflation, which will lead to different conclusions on how to mitigate or capitalize this risk. Similarly, new levels of volatility in asset markets could cause a financial company to question whether the probability of missing a dividend payment or failing to meet rating agency requirements had reached an unacceptable level.
However, incorporating the most recent data will not be sufficient on its own to ensure that the company develops appropriate long-term business strategies. In this case, incorporating a longer time horizon of historical data that encompasses a more complete set of possible events is desirable. In the case of the credit crisis, similar events occurred as recently as 1998, typified by the collapse of the Long Term Capital Management hedge fund. Then, too, an increase in the credit spreads between risk-free and risky bonds caused significant losses in a particular arbitrage trading strategy that was supposed to be risk-free. The credit crisis and “flight to quality” in the bond market at that time was caused by the Russian government defaulting on its treasury debt and came on the back of Asian financial market turmoil. It is true that stress testing of assumptions is required when recent history does not provide adequate precedents. However, in the case of the current credit crisis, an historical benchmark for performance was readily available.
Companies that are suited to withstand future crises are those with appropriate enterprise risk management practices in place. Current and timely results are required to inform appropriate management decisions. Similarly, the cars most likely to cross the finish line at a race are those capable of slowing down to meet the challenges on the road.
Jonathan Bilbul, FCIA, FCAS, is a consultant at EMB Consultancy in England. He can be contacted at firstname.lastname@example.org.