In: Finance
Practical portfolio optimization can be summarized as "finding optimal portfolio weights for available investment assets that would minimize risk while maximizing returns given investor’s risk aversion”. However, pure portfolio optimization suffers from a flaw–it gives a lot of weight to the assets (stocks, ETFs, bonds, etc) that performed well recently.
Giving lot of weightage to the assets which have performed well recently is a flaw because past performances don't guarantee future returns. In fact, practical data may show that the assets that have performed well recently, are more susceptible to poor performance in future as many assets are cyclical in nature in terms of performance of their returns. The returns on stock and gold, for example, are mostly inversely correlated. So if in the recent past, stocks have performed well (and gold poorly) then the portfolio will be more skewed towards stocks. This will make the portfolio risky as if there is a down cycle of stocks then the portfolio will not be able to take advantage of better returns from gold.
To mitigate this flaw, it is important to identify an optimum asset allocation and rebalance the portfolio periodically so that it is not skewed in favor of assets which have performed well recently. To achieve better diversification, assets selected should be least correlated with each other.