15. Portfolio management Flashcards
(112 cards)
Ex-ante risk and return measures
Calculated before the investing period gets underway
Ex-post measures of risk and reward
Calculated as time unfolds, actual values after the investment event
Ex-post standard deviation
A measure of the dispersion of returns around the average return over a given period
3 criticisms of SD
- based on past, future may be different
- measures upside and downside movements, concern on investments more on downside
- assumes upside equally likely as downside
- volatility generally not a complete measure of risk, e.g. inflation risk
3 alternative risk measures to SD
- Semi-variance: only measures returns that fall below certain value
- Shortfall probability: measures chance, not value
- Expected shortfall: measure of expected loss at a given probability level
VaR analysis
Used to estimate the capital loss on a portfolio or an individual asset over a given time period that will be exceeded with a given frequency or probability
3 features of VaR
- Time period
- Probability level (confidence)
- Loss amount or %
Historical return approach
Reorganises past returns from high to low and assumes history will repeat itself
3 ways to calibrate VaR
- Historical return approach
- Variance- co-variance method
- Monte Carlo approach
Variance- co-variance method
Assumes daily returns follow a normal (or similar) distribution
Problem with variance-co-variance method
Relies on many popular assumptions about the shape of relative frequency of losses that are unrealistic in practice
Monte Carlo approach
Mathematical model for stock price returns, running multiple hypothetical trials of the model
‘Fat tails’ problem
In real world, extreme positive and negative returns occur far more often than predicted by a normal distribution
Monthly SD of returns =
SD of daily returns x √T
VaR concept criticised since financial crisis because?
Consequently?
Many investment losses were far worse than ‘predicted’.
Consequently many VaR models have since been revisited to allow for a higher probability of extreme losses.
Stochastic modelling
A form of financial modelling of future outcomes based on ranges of values (rather than single estimates) where the value of each unknown variable, such as investment returns and inflation, is based on a statistical likelihood.
Deterministic model
Output of model is fully determined by the parameter values and the initial conditions and hence, any projections or forecasts can be made with certainty.
Stochastic models can project ___ as well as risk.
future asset or portfolio returns
Drawdown measures what 2 things?
Measures max amount an investor could have lost since the investment was at its highest price ; or
Size of loss an investor would have incurred by investing with a manager in the past
Advantage of drawdown
It refers to an empirical reality and is therefore less abstract than concepts such as volatility.
Limitations of drawdown
- longer the time series, generally the greater the drawdown (despite longevity implying robustness, experience)
- max drawdown is a single number and will therefore have a large and uncertain error distribution
Tracking error
The difference between a portfolio’s return and the benchmark or index it was meant to follow or beat
2 ways to measure tracking error
- Portfolio returns - benchmark returns
- √(Rp - Rbm)^2 / (N-1)
Name 3 factors which determine a portfolio’s tracking error
- degree to which portfolio & benchmark have securities in common
- diffs in mkt cap, timing, inv style
- diffs in asset weighings
- mgt fees, custodial fees, brokerage costs
- BM volatility
- portfolio beta