Chris Zingo, evp sales and support, Americas at SuperDerivatives, finds that independent revaluation holds the key to fair valuation of complex derivatives

We are living in a period of unintended consequences. The effort to expand home ownership has contributed to one of the largest economic collapses in market history, while the traditional Wall Street model of distributing risk - "Sell it to the other guy" - was proven to be unsustainable.

The effort to ease interest rates and create liquidity has contributed to the precipitous decline of the US dollar, while efforts to establish a common framework for OTC securities valuations has led to fire-sale asset pricing, starting with the sale of US$30.6bn of CDOs by Merrill Lynch to the Dallas-based investment firm Lone Star Funds, for US$6.7bn.

This turmoil in the financial markets has made valuation of derivative securities a front-line priority for policymakers, bankers, custodians, brokerages and everyone connected to financial markets, or trading and lending activity. Confidence in financial markets hinges on the possibility of assigning reliable, verifiable and accurate marks to diverse portfolios of complex financial instruments.

Price discovery for complex securities depends on multiple factors, such as data quality, operational and IT support, models, implementation and skill in using the models. Each of these ingredients is instrumental in producing a defendable verifiable price level.

In 2008, FASB - the panacea of a post-subprime world - established a framework for measuring fair value and expanded disclosures about fair value measurements in the form of FAS 157. Officially, it defines fair value as "an exit price; the price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants." A framework, however, is just that - a framework.

The challenge for most market participants is that there are complexities associated with OTC valuation that require interpretation, debate and plain-old good judgment; complexities that are difficult to fully define with legislation.

In September 2008, the SEC issued guidance around a number of practice issues that were the subject of much debate. It addressed issues, such as the use of broker quotes, stating that broker quotes may be an input when measuring fair value, but are not necessarily determinative if an active and indeed liquid market does not exist. It also suggested that, in weighing a broker quote as an input, an entity should place less reliance on quotes that do not reflect the result of market transactions.

This is a logical determination; however, how would the entity know whether or not the price they received was reflective of an actual transaction, a purely modelled price, or an indicative price that is no longer reflective of the actual market? It also suggests that a quoted market price in an 'active market' for an identical asset is representative of fair value.

Again, this is a logical assessment; however, how would the entity know what an 'active market' is? Guidance suggests that entities should consider the spread differentials. The challenge is that, in this market, it is difficult to identify what a rational spread differential is for an active market.

The guidance also addresses the concept of disorderly transactions and how representative they are with respect to fair value. While on the one hand it suggests that fair value should reflect a true exit price, on the other hand it suggests that liquidation sales are not orderly and should not be considered.

To contend with this grey matter, it is clear that market participants will continue to unite in the form of debate, industry events and general dialogue to share and evolve best practices that have proven successful. Market participants will continue to seek the support of third-party valuation providers to implement these best practices as the requirements (data management, operational scale, expertise) to implement industry standards continue to increase, making it virtually possible for institutions to comply on their own.

To illustrate this, let's explore two common FAS 157 best practices guidelines that have evolved as industry standards - governance and data management.

Governance
With respect to governance, we have seen a uniform adoption of the segregation of responsibilities associated with the investment/trading process and the valuation process. The verification, or shadow-processes, of the valuations department must be completely segmented from the investment/trading process.

The challenge for many institutions is that most of the expertise with respect to OTC derivatives is contained in the front office (traders, portfolio managers, quants). While most institutions - both buy- and sell-side - have segregated roles for investment/trading and valuation, they are still reliant on the judgment of their investment trading team to define fair value.

In fact, many institutions still allocate the costs associated with defining fair value to the P&L of the investment/trading group they are supporting. This is a clear conflict that must be addressed if they are to achieve true independence in terms of valuations and revaluations.

Data management
The CPA Journal, in its January 2008 publication, noted that "the fair value hierarchy, defined in FASB 157, emphasises a market-based measurement and, in doing so, stipulates a fair value hierarchy. The hierarchy is based on the type of inputs applied (the data used) to measure fair value, not the model." The journal lists the fair value hierarchy, as summarised below:

• Level 1 lies at the top of the hierarchy, where inputs are quoted prices in active markets. Level 1 inputs may be observable in markets such as the New York Stock Exchange, Nasdaq, electronic communication networks and principal-to-principal markets, where prices are negotiated independently between the parties with no intermediary.
• Level 2 inputs are in the middle of the hierarchy, where data are adjusted from similar items traded in active markets, or from identical or similar items in markets that are not active. Level 2 inputs do not stem directly from quoted prices.
• Level 3 inputs are unobservable and generated by the entity itself. An asset retirement obligation for an oil well, for example, would include expected risk-adjusted cashflows, using the company's own data.

The focus of much debate in terms of valuations is around the classification of Level 2 and what constitutes a Level 2 asset verses a Level 3 asset. The values associated with Level 2 assets come from 'observable inputs', meaning they are gathered from sources other than the reporting company and that they are expected to reflect assumptions made by market participants.

Unlike Level 1, where the quoted prices are observable in the listed markets, the OTC markets are very distributed with respect to where trading occurs. Observable data can come from a variety of sources: the broker marketplace, the dealer community, third-party pricing providers and exchange publications.

In addition, within the sources themselves, the trading is very fragmented. Many of the sources' visibility into price discovery is limited to the 'flow' they observe in their trading operations.

Like most participants, they are forced to make assumptions about assets they don't trade or price daily. These assumptions can come from different interpolation methods, survey methods, historical comparisons, anecdotal information and traditional 'best-guess' estimates. As a result, differences in opinion about the same asset can be dramatic, especially in a volatile market environment.

The following example illustrates the common variances that exist between sources (contributions from two different sources for INR FX volatilities at a similar time in the same day).

Notice the three-month tenor. There is a 16% difference in the quoted market.

As an entity looking for observable inputs, what should you do? Do you average the two quotes? Do you assume one is correct and the other is incorrect? Do you assume both are incorrect?

The answer could be any one of the options listed above. You cannot determine the answer without truly understanding the nature of the source, consistency of the source and the competencies of the source. Furthermore, given the volumes, it becomes increasingly difficult to understand where the market truly is unless you employ quantitative procedures to check your data.

For example, the illustration below represents the term structure submissions from the previous sources. The jagged points from the contributor on the left should immediately result in a 'red flag' in your process. Jagged points are quantitative signs of potential model error or input bias.

Given the variance of the markets in the first illustration coupled with the quantitative red-flags in the second illustration, this currency pair should be flagged for immediate investigation. A manual intervention should then occur to assess and decide. This may involve communicating with both sources and/or checking additional sources.

Conclusion
It is becoming clear that institutions - both buy- and sell-side - accountants and regulators must seek independent valuation services that provide wider coverage and deliver independent price opinions reflective of the current market, backed by verified data, models and technology. Transparency facilitated by fair value accounting instills confidence in investors and in the financial system as a whole. Combine this with solid risk management practices and the route to the recovery of confidence is laid out.

Pricing derivatives at the point of revaluation is an art, underpinned by some serious science provided by independent and specialised third-party valuation companies, rather than a single bank source. This technique will meet the requirements for fair valuation of financial instruments, which I believe will become required for all money managers. The independent valuers will be the partners of the banks and funds, in the same way the auditors are the necessary partners of public companies.