Building a better model: Enron's downfall should alert directors and officers writers that their underwriting models must be accurate, up to date and effective for their entire policy portfolio. (Property/Casualty: Directors and Officers).

In many cases, selection and pricing of risk can be only as good as the model the insurer uses to do the job. Developing valid models means insurers must carefully assess relevant factors to ensure that pricing, terms and conditions accurately reflect the exposure. The risk that the model might fail is a real concern for insurers, who must realize that the shelf life of a model will vary depending on the volatility of risk exposures and unforeseen conditions. Not only must insurers update their models to accurately reflect changing exposure for each individual risk, they also must modify the models so they take into consideration the larger policy portfolio.

One of the potential problems of risk selection is that it relies significantly on the accuracy and quality of the financial reports filed by publicly traded companies. When either the accuracy or quality of those reports is deficient the formal model is doomed to fail. Hence, the principle of transparency -- full, fair and complete disclosure--is critical for investors and underwriters. The lack of transparency in financial statements of a few notable companies has created a crisis of confidence in financial reporting. When transparency is absent or illusory, the cumulative effect of model failures may result in system breaks such as the Enron meltdown, a watershed event that doubtlessly will result in changes in the federal securities laws and public accounting standards. Lawmakers have responded by enacting corporate reform legislation. On July 30, 2002, President Bush signed into law the Sarbanes-Oxley Act of 2002, which is an important step in protecting investors and restoring public trust.

Texas-based energy trader Enron Corp. filed for federal bankruptcy protection on Dec. 2, 2001, exposing insurers to losses of more than $2 billion. The ripple effect of Enron is already being realized across the insurance industry For example, one of Enron's insurance coverages that has been frequently mentioned in the press is directors and officers liability. Indeed, Enron officers and directors are looking to their D&O policies as a source of defense payments. The potential multimillion-dollar loss that may be covered by Enron's D&O carriers, however, is significantly lower than Enron's actual exposure that, on a relative basis, rivals a categorical catastrophe loss.

Learning From Enron

In light of these events, D&O insurers should consider Enron Corp. as a case study and recognize that avoiding the herd mentality requires independent analysis, technical acumen, and the discipline to underwrite within self-imposed boundaries. In October 2000, Chief Executive magazine selected the Enron board as one of the five best in the United States. And in 1999, Enron's former chief financial officer, Andrew Fastow, was a recipient of CFO magazine's Excellence for Capital Structure Management award. Beneath a glimmering facade, six red flags revealing Enron as a high-hazard risk were present prior to Enron's bankruptcy filing: complex accounting, hidden liabilities, related-party transactions involving senior officers, high-risk industry sector, numerous subsidiaries and off-balance-sheet transactions.

In addition, the risk profile of Enron Corp. appears to be more like that of an unregulated broker/dealer than a pipeline/energy/utility company. Broker/dealers have generally been considered a relatively higher risk group for D&O insurers.

Therefore, prudent underwriting analysis requires more than the status of Enron as number seven on the Fortune 500, numerous strong "buy" ratings by top-tier sell-side analysts, high-grade ratings by credit analysts, audited financial statements by Arthur Andersen, and a well-regarded and respected board of directors. A firm's "capitalized reputation" can rapidly erode during times of uncertainty, particularly when price-earnings ratios are lofty. Therefore, underwriters must rely upon proven and time-tested analysis, not Wall Street's "flavor-of-the-month" metrics.

The question is: Does the underwriting analysis truly identify a company such as Enron as a higher risk relative to the overall policy portfolio? This accuracy can be measured by the limits of liability, terms, conditions, and pricing offered by the D&O underwriters. If the company is identified as higher risk relative to the overall portfolio, then the risk-selection process for identifying exposure is valid.

Model failure occurs in this case if the risk is rated as low to medium hazard. If the potential for catastrophic losses considered as outliers exists, and the overall portfolio includes such accounts, then the premium-pricing model ought to contain loadings for such a scenario. If so, then the premiums collected for the entire portfolio should be substantial enough to absorb such shock losses over a multiyear period. If massive exposures are truly categorized as an outlier, then a lower frequency for such exposures must be reflected in the portfolio.

Loss causation, severity, frequency and variance to the forecast are variables that should be identified and measured. Insurers should stress test their book of business to ensure the risk exposure has not changed and that balance is present within their portfolio. An insurer's diversified portfolio mitigates the risk of losses occurring from model failure.

The pricing model for exposures needs to be based upon accurate and empirical methods. Such analysis should assess the likelihood of meltdowns similar to Enron. Analysis of losses should include not only insured losses, but also total liability had the insured purchased sufficient limits of liability. Most models exclude such losses as outliers and state that inclusion of the loss is not relevant. The argument that these approaches skew results is valid under certain circumstances. Insurers need to address risk selection and underwriting models to reflect the changes in the environment, such as regulatory oversight or deregulation, judicial decisions, legislation, stock market volatility and claims inflation.

A model also evaluates variables that include risk, exposure, experience and exogenous load factors. It translates them into a qualitative picture of risk management. The model should be used for the entire portfolio of policies. If the model is not current or correct when realistic assumptions are taken into account, then the model needs to be changed.

Risk of Model Failure

The importance of the model s risk of failure is increasing with the greater level of uncertainty and volatility Sept. 11 brought to reality the unimaginable. The collapse of Enron defied conventional wisdom. If the insurer's objective is to write homogenous, preferred risks with predictable underwriting results, then the underwriting risk-selection model should accurately identify the likelihood of loss. Risks that are not quantifiable or predictable within a normal range are difficult to assess. The inability to identify expected risk and subsequently obtain an expected return translates to exposure that the market may not be able to price.

The consequences of model failure are financial losses, unpredictable underwriting results, lack of confidence in underwriting decisions, pricing gaps and ultimately, market dislocation. The lack of confidence in underwriting should not be understated, as the countervailing theory of "Enronitis" is the cockroach theory--if one is found in the kitchen, there are probably more in the basement. Confidence in underwriting decision making needs to be earned, coveted and protected. Continuation along the path without a sober introspection that elicits meaningful constructive change will lead to disequilibrium. Expectations for an improved variance to probability of loss are contingent upon aggressive action by insurers regarding nonperforming operations and updating of risk-selection and pricing models to reflect changes in risk exposure.

Insurers should be aggressively proactive in implementing constructive changes in the underwriting decision process. The results of change should be identifiable and measurable. The efficacy of such changes should be reflected in greater underwriting profitability.

One tool to ensure homeostatic control of operations is the audit. An underwriting audit should reveal process, procedure, underwriting discipline, adherence to guidelines, decision making and the accuracy of portfolio mix. The audit of claims is equally or, in some cases, more important than the underwriting audit. It is where reality meets theory. A claims audit should enable an insurer or reinsurer to acquire a better understanding of process, procedures, claims management, efficient resolution, reasonable settlements, reserving practices and prudent tolerance to risk.

One measure of effective insurance operations is the interworking of underwriting and claims. Underwriters should meet with claims regularly to ensure that underwriting decisions reflect claims trends in exposure. A claims audit is a valuable tool to aid in the efficacy of the underwriting decision-making process. Management and their reinsurers should more closely scrutinize underwriters that inadequately price their products relative to the risk. Heightened frequency of claims that are piercing through the excess-of-loss reinsurance ought to be examined. Underwriting files for policies that are the subject of significant losses need to be reviewed. The analysis of the underwriting decision making and thought process should be documented. The assessment of risk and adherence to underwriting guidelines should be clearly defined. The rating plan, underwriting work sheet, terms and conditions need to accurately reflect the risk. Such evaluation of pricing, terms and conditions ought to be measured on a relativ e risk basis to the overall policy portfolio and industry benchmarks.

RELATED ARTICLE: 7 Steps to Better Models

Insurers can minimize or avoid the risk that their selection and pricing model will fail by implementing the following:

* Underwrite perceptively.

* Stay on top of changes in risk exposure.

* Incorporate changes in legal, economic and regulatory environments.

* Stress test the portfolio. Use pricing, sensitivity and volatility analyses. Implement frequency/severity claims simulation methods.

* Categorize risk accurately according to meaningful metrics, instead of Wall Street's hype.

* Make sure the underwriting and claims departments are talking to each other.

* Conduct frequent in-depth audits of underwriting and claims.

Christopher Duca is president of Navigators Pro, a division of Navigators Management Co., and a member of the board of directors of Navigators Insurance Co. He is based in New York.