Difference between revisions of "Risk Management Wisdom"
From Open Risk Manual
Wiki admin (talk | contribs) (Created page with "== Risk Management Wisdom == === Complexity === <ul> <li>Identifying, quantifying and managing complexity is one of the least developed areas in financial riskmanagement <li>...") |
(No difference)
|
Revision as of 13:53, 11 September 2019
Risk Management Wisdom
Complexity
- Identifying, quantifying and managing complexity is one of the least developed areas in financial riskmanagement
- In the financial universe there are no isolated systems: personal, corporate and country risks interlinked in a correlation web
- The web of financial contracts and interdependencies is complicated but finite. in due course it can be understood and tamed
- Financial risk model complexity and quality are not necessarily correlated. But complexity is also in the eye of the beholder
More Wisdom
- Applying quantitative models to financial phenomena is subject to severe misspecification, whether intentional or unintentional
- After the simplest binary, risk on / off choice, next step up in risk management complexity is the traffic light system: green, orange, red
- Reputation in banking is not optional but a license to operate. Reputation risk introduces a zero tolerance approach
- When there is no empirical basis to support risk assessment the fallback is polling expert opinion. Human experts will use analogy
- When you set a limit on trading you create anchoring bias. Read Kahneman
- Controlling dynamically the amount of risk exposure by setting and updating risk limits is a standard tool for risk management
- First recorded incident of bank risk concentration: Florentine banks' Bardi and Peruzzi sovereign exposure to King Edward III
- Regulation is supposed to prevent financial crises. the track record seems not good but we don't really know a world without it
- Financial crises are recurring economic phenomena where prevailing contractual arrangements / designs get severely challenged
- Scenario analysis is a basic building block for a probabilistic approach to risk
- Diversification versus model risk are the two key phenomena relevant for managing portfolio risks
- In a crisis, all correlations go to one. Except this phenomenon concerns market expectations and prices and actual events might be less correlated
- Risk insurance can remove risk but does create new risks (exposure to the insurance provider and other residual risks)
- There is risk and there is the perception of risk. The two are only loosely correlated and perception is generally more volatile
- One of the simplest and most effective forms of risk management is risk avoidance (also termed zero risk appetite)
- Operational risks definitely scale with the size and complexity of the organization but the precise law is not easy to establish
- Some risks can be usefully decomposed into a probability of occurrence and a severity of occurrence. Others have a continuous range
- Risk is about the deviation /uncertainty around expectation. A bad outcome expectation with little remaining uncertainty is not risky
- There is no physical law to prevent even the deepest, most active, market to become illiquid
- Risk is not linear. Robust ratings need log-scale: 10%, 1%, 0.1% hence limited in number. Alphabet soup not a good risk metaphor
- Credit risk analysis looks at the 1) ability and 2) willingness of an entity to fulfill its contracts. It gets messier very fast
- All risk models built on past experience. some may contain an element of truth about the future
- All forms of financial risk are subordinate to sovereign risk