Can anyone out there explain " Optimisation" as an approach to constructing an indexed portfolio in simple english?
would highly appreciate the assistance.
Optimization focuses on systematic sources of return . If you take a simple 3 factor model a la Fama , then the main drivers are Market sensitivity , Size effect , value effect. An indexer that uses optimization would likely pick stocks by matching these sensitivities closely ( i.e. trying not to deviate more than a small percentage of standard deviation from the benchmark returns for each factor )
Carhart adds momentum although Fama calls it an anamoly ( i.e. non systematic source )
You could construct a factor model with other factors inclused such as liquidity , quality ( debt to capitalization )or other fundamental leanings of the benchmark.
Basically you regress the portfolio returns to each factor return in the benchmark ( constructed in a specific way using benchmark stocks and weights ) in a multiple regression to get the sensitivities ( i.e. betas ) for each expsoure .
For example for Market sensitivity ( or beta ) you simply regress benchmark return to the market return .
For Value sensitivity construct a factor portfolio return as a difference between average return of all high value ( high book to market ) minus the average of all low value ( low book to market ) ( Fama mixes size into this by taking the average as an average of two size portfolios ( big stocks and Small stocks ) in each value category )
The book also mentions that deviations of the portfolio are measured as number of standard deviation units from the benchmark. So the control excercised is on this level of deviation