Sunday, January 21, 2018   You are here:  Features   Search
  Industry News Minimize
  Non-Conventional Wisdom: Debunking Myths in Risk Management
Non-Conventional Wisdom: Debunking Myths in Risk Management

By Richard Neiman

In the aftermath of the subprime mortgage meltdown, risk managers and regulators around the world are doing a post-mortem.  And the diagnosis is ironic. In some cases, the very practices designed to contain the risks of subprime or non-conventional lending actually compounded the problem.
The subprime crisis has exposed the weakness of underlying assumptions in risk management. These assumptions were the prevailing conventional wisdom, and they are reasonable on the surface. So reasonable, in fact, that they went largely unchallenged until inconvenient realities revealed their shortcomings.
I encourage you to read the many outstanding reports on the subprime crisis, such as the Senior Supervisors Group study issued in March, and the Financial Stability Forum report and recommendations released in April. These reports explore the assumptions and resulting lessons learned in detail. I hope you take their observations to heart, and find creative ways to apply them to your own business context.
To help get you started, here is my top list of assumptions that were tried and found wanting.

Assumption #1:
Securitization effectively disperses risk
Now, it is true that securitization partially disperses risk and certainly no one wants to concentrate risk. But the assumption that securitization allows an originator to wash its hands of a risky deal has proven to be patently false. Repurchase demands have boomeranged back onto lenders. And mortgage bankers in particular have sometimes found themselves with insufficient repurchase reserves, and were pressed into bankruptcy.
And there was the unintended side effect of complacency. When risk is disaggregated, no one in the distribution chain has the incentive to invest in sufficiently robust risk containment. One particularly vicious form of complacency undermined the quality of underwriting standards.
When the risk of default was assumed to be passed on immediately, credit standards and consumers suffered. The widespread disregard of the borrowers’ ability to pay is a key symptom. There is a marked contrast here with the dynamics of traditional banking, which looks to establish an ongoing relationship with the consumer and a commitment to the community. Instead, misaligned incentives drove a wedge between originators’ short-term and consumers’ long-term best interests.
This isn’t to suggest that securitization is an inherently flawed model. On the contrary, securitization as a process is positive: it increases liquidity and access to the capital markets. But we need to think much more holistically when evaluating residual risk, contingent responsibilities, and fractional interests.

Assumption #2:
Compartmentalized risk management increases protection
The industry was certainly aware that increased risks accompanied the gains of securitization. And that re-securitization, with its host of ever more complex structured products, was a subject for specialists. It was logical, therefore, that a more compartmentalized approach to risk management would develop. Credit risk, market risk, and operational risk are disciplines in their own right, and demand unique expertise. But these specialties should be kept within an integrated approach, and not become silos.
The interrelationships of complex products meant that the whole risk was more than the sum of its parts. And that a wide-angle view is a needed complement. There was a tendency to focus on credit risk at the expense of market risk, or vice versa. For example, loans held for sale are typically a market risk. But they can be quickly transformed into a credit risk if the market dries up and those loans unexpectedly linger in portfolio.
Assumption #3:
Quantitative assessments are highly reliable
Statistical modeling is an important tool in monitoring those credit and market exposures. There are many benefits to integrating a quantitative element into the risk management process. But models do not replace or trump critical qualitative assessments, and in the heyday of the subprime splurge there was an over-reliance on quantitative results.
I’m reminded of the old “garbage-in/garbage-out” principle of computing. The end result is only as reliable as the underlying data and parameters used. Data from more benign economic times was feeding these models, as well as volatility projections for new products without a long price history. This may have tipped the scales on the models, and reduced their usefulness in identifying the risk of future market shocks. And even the most finely tuned models are not designed to capture the risk of remote but catastrophic loss.

Assumption #4:
Ratings scales for different instruments
are comparable
In a world of such complex products and analytical methods, it can be comforting to rely on the evaluation of a third-party. And so a final point that I would like to make is this: there was a widespread assumption that credit ratings agencies had access to superior inside information on otherwise opaque transactions. And therefore their assignment of the same AAA rating to CDOs as to corporate bonds was accepted less critically.
It turns out that the ratings agencies had access to far less unique data than supposed. And that CDOs, being a layered product, are inherently more risky than traditional debt instruments. The AAA rating for both may have lulled unwary investors who didn’t read deeper into the prospectus. It’s a bit like a shopper skimming those government food labels, and then assuming that 100 calories always have the same nutritional value. The outward label may have the same final number, but the ingredients can have very different properties.
I’ve really just scratched the surface, in considering the lessons learned from the subprime crisis for risk managers. And every day new insights are uncovered. I think that an important dialogue is underway among industry and regulators. We have a common goal to identify and mitigate risk, both at an institutional and a systemic level.
Conventional wisdom may not have been up to the challenges of non-conventional lending. But after we debunk the myths, I’m confident that we can develop new strategies, or “non-conventional wisdom” to retain the benefits of innovation, while safeguarding against similar market disruptions in the future.

Richard Neiman, superintendent of the New YorkState Banking Department, writes on regulatory issues for Banking New York.

Posted on Thursday, July 03, 2008 (Archive on Wednesday, October 01, 2008)
Posted by Scott  Contributed by Scott


Current Rating:

Privacy Statement   Terms Of Use   Copyright 2013 The Warren Group    Login