Brooks has a good one today (just read the bold stuff):
Over the past decades, we’ve come to depend on an ever-expanding array of intricate high-tech systems. These hardware and software systems are the guts of financial markets, energy exploration, space exploration, air travel, defense programs and modern production plants.
These systems, which allow us to live as well as we do, are too complex for any single person to understand. Yet every day, individuals are asked to monitor the health of these networks, weigh the risks of a system failure and take appropriate measures to reduce those risks.
If there is one thing we’ve learned, it is that humans are not great at measuring and responding to risk when placed in situations too complicated to understand.
I heard a story recently about a group of insurance executives meeting with BP’s risk manager in the 90s. She reiterated a common refrain: we don’t need insurance because we have a huge balance sheet.
And yet, insurance still has something to offer:
1. Individuals often overstate risk tolerance, especially with respect to organizations. I think that this is a combination of Robin Hanson’s Near vs Far bias and agency cost. People rely too much on silly capital models and not enough on anticipating the response of those individuals who (collectively) actually pay the costs (shareholders, stakeholders, etc) of risk when it goes pear shaped.
2. A market based appraisal of what the risk is worth. Sure, insurance underwriters may not understand it perfectly, but at least get a second opinion. You might miss something and the costs of that are, obviously, huge.
In spite of the fact that most risk is fundamentally incomprehensible, the insurance industry is actually quite stable. It continues to baffle me, to be honest, but I can’t argue with success.