An investment portfolio can be based on an investor’s risk tolerance, age or lifestyle objectives. Target weightings to various asset classes, like shares, are then used to build that portfolio and generate a desired return but this approach has flaws. This is because target weightings are often built on historic average returns.

Typically, there will be a range of target weightings to different defensive assets (for example, cash and fixed income) and growth assets (for example, shares and property) to deliver a maximum expected return from a portfolio based on a given level of risk. This asset allocation process will often be done using a process called mean-variance optimisation. This involves taking the historical average or “mean” return from each asset class and the volatility or “variance” of the returns, as well as how returns from one asset class vary with the returns from others (known as the “covariance”). The optimisation process then finds the optimal weights to each asset class.

While mathematical asset allocation techniques may look good on paper, they are backward-looking. This is because the optimiser favours a higher weighting to asset classes that have performed well historically, which can be dangerous. For example, Australian bonds returned 7% per annum over the past 20 years with relatively low volatility (3% per annum) yet Australian government bond yields are currently less than 3% per annum so the historical average of 7% is probably not a reliable estimate of future returns. To prevent the portfolio becoming overweight in asset classes which performed well historically requires subjectivity. We can replace historical returns with long-term forecast returns in the optimisation process but forecasting is difficult.

Another technique to account for current conditions is to still use historical returns for determining long-term average weightings to each asset class but then to periodically recommend investors adjust asset class weightings away from long-run targets based on how we expect the asset class to perform over the next few years.

August asset allocation changes

Our asset allocation has changed this month (August). We are now neutral in cash, underweight fixed income and overweight alternatives. We are still recommending a neutral holding in international equities and prefer unhedged currency exposure. Our underweight exposure to Australian shares is unchanged, as is our neutral holding in property.

Our asset allocation summary:

• Cash: Move to neutral position. No preference between term deposits and at-call cash.

• Fixed income: We have reversed our overweight position in fixed income and have shifted this capital into alternative investments so move to underweight in fixed income. Developed world government bonds are expensive and offer poor absolute value, so we prefer products with limited interest rate risk. We suggest an equal split between Australian and (hedged) international bonds.

• Australian equities:Remain underweight. Growth outlook is lower than other markets and valuations are above fair value. Favour selected industrials such as financials, healthcare and utilities.

• International equities: Given higher valuations in developed market shares, hold a neutral weighting. Maintain unhedged currency exposure for now. Favour Europe and Japan over the United States. Emerging markets are cheaper than developed markets so maintain exposure and favour Asia over other regions.

• Alternatives: Move to an overweight allocation. Alternatives are part of an overall strategy of building allocation to assets with a low/moderate correlation to equities.

• Property: Hold a neutral allocation to commercial property. At current pricing, Australian and international property appears fair value.