Risk Management at Banks
I have never been a fan of VAR (Value at Risk), but I recognize that mathematical techniques are only as good as those that use them. Questions arise with any quantitative risk technology:
- What’s my time horizon? (What’s your longest asset or liability?)
- Do I have to be good over this entire time horizon, or just the end? (The whole thing, sorry.)
- How do I work with options in assets or liabilities? (Assume optimal exercise by the option holder.)
- What are the worst losses can I take from this business activity? (Much worse than you can assume, and this present crisis is an example of that.)
- How do I model liquidity of liabilities? (Assume they exit when it is in their interest.)
With one employer, he invited me to consult for the Asset-Liability committee of his bank. Having been a risk manager inside two life insurance companies, when I reviewed the documents, I was surprised, because they were so much less sophisticated than what life companies of a similar size did.
With banks, the grand weakness is in the assets, and the analysis should focus on two things:
- Liquidity of the assets versus liquidity of the liabilities.
- Potential credit losses of the assets versus the surplus of the bank.
I write this because of the commentary of Taleb and Bookstaber. They are bright men, but they have never managed the risks of a financial institution. Leverage ratios are not enough. One must dig into the loss experience and analyze whether emerging losses might overwhelm the capital of the institution. One must also look at risk-based liquidity — what is the likelihood of running out of cash?
There is always a tension between rules versus principles. What must first be admitted is that both can be fuddled by sinful men. Rules can be observed, and cheaters bring items that meet the letter of rules, but violate the spirit of the rules. Principles can be bent by those that implement the principles. Neither is a perfect solution — better to settle on one way of regulating, though, and understanding the soft spots, than to vacillate.
Perhaps the banks need to employ actuaries. I don’t say that so that friends might find work, but because many banks do not get how to preserve their existence. Actuaries think longer-term; they think about scenarios where loss experience might prove to be unsustainable. They are more skeptical on risk compared to most bankers.
With that, I commend all who read this to be careful, and to analyze financial situations carefully. Don’t follow the crowd.
September 12th, 20093:10 pm at
Section V.B of the Senior Supervisor’s report “Observations on Risk Management Practices During the Recent market Turmoil” recommends the use of multiple measures of risk, including 1) notional measures, and 2) both conditional and unconditional measures of risk.
This discussion can be found at http://www.newyorkfed.org/newsevents/news/banking/2008/SSG_Risk_Mgt_doc_final.pdf
Paying attention to simple notional measures of risk helps compensate for the potential mis-rating of exposures, as firms will usually be willing to hold much larger volumes of highly-rated credits. In terms of tail events, “fallen angels”, whether corporate or mis-rated super-senior tranches, usually contribute to the largest realized losses. Notional measures compensate for the one-way risk of highly-rated exposures.