Stories Below come from our Media Partner 3BL Media - Click their logo or any of the stories for more information
February 15, 2008 11:50 AM Age: 9 yrs
Model Risk and AccountabilityCategory: Acc Commentary & Opinion, CG Commentary & Opinion, A/F Commentary & Opinion, Eleanor Bloxham
Source: Eleanor Bloxham
“Model risk” has been getting a lot of press lately in the midst of the credit meltdowns. In brief, model risk is the risk that the model [that] one is relying on to measure risk (or anything else for that matter) is invalid. The reasons why a model may be invalid include the fact that the model is poorly constructed and does not reflect all important drivers, uses out-of-date data or assumptions, inaccurately models the relationships between important elements, or, is used in situations where it does not properly apply.
In the risk area -- which is all the rage right now -- there is a great deal of groupthink that goes into risk modeling. For many years, I have spoken at and chaired risk conferences. (One that I remember at the World Trade Center included the push by some of the speakers of “securitization” as a risk management tool.)
The groupthink for years was that models were something you could learn how to put in place following what everyone else was doing.
While the solution may be (in part) to have math types involved, the problem of model risk also is because many modelers are mathematicians who tend to act as if the model were a simple problem stated in school. “Let’s assume that….what would be our outcome?”
The problem in many models is the “let’s assume” part which is never re-evaluated and in fact may have represented faulty logic in the first place.
When I’ve questioned modelers pointedly about inaccurate models, they have said to me: Well, the assumptions may not exactly be true but “the math is easier this way.”
For example, in criticizing and proffering solutions to the state of risk models as I do in my book Economic Value Management, I point out, for example, limitations in the standard models of credit risk assessment accepted by the groupthink community today. The book shows how making some of the commonly accepted assumptions can create clear distortions. (click here for information on the book)
For example, for credit risk, typical models are constructed of the “let’s assume that the probability of default can be modeled thusly.” The problem is that this “group think” has caused most modelers to forget over time the “let’s assume” shortcut. This has resulted in a failure to properly assess the model’s validity and to appropriately modify the model -- or at least to monitor and measure the impact and distortions that the math shortcut entails.
Away from this [present] scary credit world, we see the same “model risk” occurring in many corporations with respect to the use of accounting numbers. Those accounting numbers represent “the accounting model.” And the accounting model is based on rules or principles of reporting designed by the accounting standard setters with a set of ends in mind, which have varied over time.
While much of the focus inside corporations in the post-Enron era has been on “are we reporting the numbers accurately in accordance with the standards,” the model risk issue that is too often ignored is how these numbers are used.
Just as the mathematicians may find it easier to use overly simplified math to model complex situations, so too, many boards of directors and corporate management teams find it convenient to use accounting net income for a variety of purposes -- even though it is not always the right model for the job.
Too frequently it becomes a handy way of best measuring results or rewarding CEOs. This misuse of the model represents a dangerous example of model risk as CEOs use this model to drive their awareness of results and drive their strategies – and seek to meet the requirements of the board by taking non-economic decisions because those decisions promote the results in an accounting model that is being misused by the board as a tool of accountability.
The fix to this important accountability problem will come in a number of forms -- as will the corrections to risk models more general. It will begin with awareness by more boards, CEOs and investors in recognizing these model risks and acting to establish and encourage models that provide a better way of measuring results, driving strategy and of rewarding CEOs.
Alex Lajoux, co-author of the “Art of M&A Financing and Refinancing,” wrote in reviewing Economic Value Management that it was “an authoritative guide for corporate leaders who want to break out of financial ratio boxes to build sustainable value." As the quote implies, stepping out of “group think” to think more carefully about models and accountability will require corporate and board leaders with the wisdom and courage to break away from invalid models and move to models that will actually serve them.
# # #
Eleanor Bloxham is a strategic governance and valuation authority, author and advisor, and CEO of The Value Alliance (www.thevaluealliance.com) and Corporate Governance Alliance. She welcomes your comments at firstname.lastname@example.org
Comments Copyright © 2008 The Value Alliance Company. All rights reserved17297 times viewed
|HOME | ABOUT THE SITE | REGISTRATION INFORMATION | VOICES: FEATURED COMMENTATORS AND BLOGGERS | SPECIAL SECTIONS|