Safe, diverse indexes

What do we define as “sufficient” diversification? Enough to cover all industries? You would need sufficient to cover at least each GICS (Global Industry Classification Standard '” or sub-group of the ASX). That's 24 stocks .

Which stocks do you choose and how do you weight them in your portfolio: equal weighted? cap-weighted? If you are seeking to generate benchmark-like returns, the problem that arises with fewer stocks is the level of active portfolio risk (risk relative to the benchmark). There is a danger in looking at the total risk numbers of the portfolio and the benchmark and thinking “that's not a large difference”, but in fact it is.

For a 20-stock portfolio an active risk value (that is deviation from benchmark) of 6–9% would not be uncommon, although it would be very aggressive for a typical active manager. Yet the total risk of this portfolio is not too different to the benchmark. This means that there exists the chance that the portfolio's return will differ compared to the benchmark by more than 9% on average one year in three. If you accumulate that difference over time, the distribution of ultimate portfolio values can be very large. Yet Christoph Schnelle’s argument (see Index funds: why bother?) is based on the assertion that the difference in total risk numbers are not large.

If we are asking people to evaluate whether to hold a reasonably diverse number of stocks or an index fund, you may reach “reasonable” diversification with 20 stocks. The biggest risk by doing this, however, is event risk. If one of the names that you own goes bust, that's 5% loss to the portfolio (12.5% for an eight-stock portfolio) '” is that something you're willing to risk as an investor?

Also if you choose to own the securities in equal weight, how frequently do you rebalance the portfolio? This would be necessary as the performance of each security will differ through time and hence the security weights will change '” this has implications for realising capital gains.

Even if you hold them in market cap proportions, you would still need to rebalance to stay reasonably neutral to the industry weights and this doesn't even take into account the risks that arise from other factors, such as size; For example, small cap effects are real, but cyclical and even a well-diversified portfolio may not capture this characteristic.

The benefits of an index fund over trying to achieve a diverse portfolio directly are:

• Mitigates event risk. Sure we've had a few One.Tels and HIH's in the portfolio, but the level of diversification is such that the portfolio effect is negligible. As the security falls in price, the marginal dollar invested into the portfolio buys proportionately less of the stock and with Vanguard’s optimised approach gets to a point fairly quickly where it would not be purchased

• Reduces the need for regular portfolio rebalancing. Indexes do change, and need to change through time to reflect the evolving market structure. Index funds can efficiently capture that evolution with less trading activity (with the flow-on benefit of improved after-tax returns).

• Positive capital flow into the fund mitigates unnecessary portfolio turnover. The benefit of positive capital flow means that we don't have to turn over the portfolio as much as implied by the benchmark turnover to be able to capture the market return.

Roger McIntosh is head of global equities at Vanguard Investments Australia.