Most of the risks that concern us exist in systems. In massively complex systems.
However, our approach to risk assessment is often to isolate certain risk/loss events and treat them totally marginally. That works fine when the events are actually marginal to the system but it may put us in a worse situation if the event triggers a cascading failure.
Within a system cycles are found. Cycles that can ebb and flow over a long time. And cycles that are self dampening or cycles that are self reinforcing.
The classic epidemiological disease model is an example of a self dampening system. The dampening is caused by the fact that disease spread is self limiting. Any person will have so many contacts with other people that might be sufficient to spread a disease were they infected. In most disease situations, the spread of the disease starts to wane when enough people have already been exposed to the disease and developed immunity so that a significant fraction of the contacts that a newly infected and contagious person might have are already immune. This produces the “S” curve of a disease. See The Dynamics of SARS: Plotting the Risk of Epidemic Disasters.
The behavior of a financial markets in a large loss situation is a self reinforcing cycle. Losses cause institutional investors to lose capital and because of their gearing the loss of capital triggers the need to reduce exposures which means selling into a falling market resulting in more losses. Often the only cure is to close the market and hope that some exogenous event changes something.
These cascading situations are why the “tail” events are so terribly large compared to the ordinary events.
Each system has its own tipping point. Your risk appetite should reflect how much you know about the tipping point of the system that each of your risks exists in.
And if you do not know the tipping point…