Risk Over Time
Time Horizon is the expected length of time an
investor allows in order to meet financial goals.
There was (and still is) some confusion in the finance community (academics
as well as practitioners.) about how the time horizon affects risk. The short
answer to this is that it depends on the definition of risk used: In a "Markowitz-type"
of world where the standard deviation of returns is the relevant risk measure,
time horizon does not affect risk. If other risk measures are
used (for example, Roy's probability of not achieving a certain minimum return),
then substantial time horizon effects exist.
Very often, only the end-of-period results (=returns) are discussed, without
paying attention to sub period results and their sequence (=path
dependency is ignored). This can be fatal in real-world application
since illiquidity and bankruptcy define certain maximum losses from where no
recovery is possible.
The sampling frequency at which risk is measured is also very important. For example, if returns are not independent over time, annual risk figures cannot be calculated easily from monthly data.
|
