Growth Optimal Investment Strategy Efficacy: An Application on Long Run Australian Equity Data

A number of investment strategies designed to maximise portfolio growth are tested on a long run Australian equity data set. The application of these growth optimal portfolio techniques produces impressive rates of growth, despite the fact that the assumptions of normality and stability that underlie the growth optimal model are shown to be inconsistent with the data. Growth optimal portfolios are constructed by rebalancing the portfolio weights of 25 Australian listed companies each month with the aim of maximising portfolio growth. These portfolios are shown to produce growth rates that are up to twice those of the benchmark, equally weighted, minimum variance and 15% drift portfolios. The key to the success of the classic, no short-sales, growth optimal portfolio strategy lies in its ability to select for portfolio inclusion a small number of Australian stocks during their high growth periods. The study introduces a variant of ridge regression to form the basis of one of the growth focussed investment strategies. The ridge growth optimal technique overcomes the problem of numerically unstable portfolio weights that dogs the formation of short-sales allowed growth portfolios. For the short sales not allowed growth portfolio, the use of the ridge estimator produces increased asset diversification in the growth portfolio, while at the same time reducing the amount of portfolio adjustment required in rebalancing the growth portfolio from period to period.