Emerging markets are inefficient, almost by definition, and you cannot afford to deal with any choice affecting them with the same emotionless rationality as developed markets.

Morgan Stanley, the IFC and ING Barings are probably the only providers of consistent equity indices for these markets. Each provides a broad benchmark that captures aggregate emerging markets performance, and a narrower 'investable' index that approximates the foreign investor's opportunity.

All these indices are weighted by market capitalisation, and this scheme enjoys popularity for good reason. In portfolio or index construction terms, the weights are self-scaling. This means that when the assets move one relative to the other, the index weights change in the same way.

As an intellectual rationale, believers in Capital Asset Pricing Model will know (whether or not anyone else agrees) that the market-capitalisation weighted portfolio is the only optimal way to hold risky assets. This assumes that event risk is being priced efficiently. If not, we lose our theoretical basis and capitalisation-weighted baskets may be inefficient portfolios and inappropriate benchmarks.

Equal-weighting confers on every asset in the index the same weight. Equally weighting countries in a portfolio requires you to readjust their weights as relative return differences drive them apart. One of the keys in understanding successful emerging-market investment is having the courage to assert that equal-weighting minimises event risk. Many of these countries are still experimenting with the use of free markets in making economic decisions. Starts and stops in government intervention, and a simple lack of stabilising institutions, create enormous return volatility.

But most of this risk is political, and political risk tends to be local. So this high volatility is diversifiable. A capitalisation-weighted index is less diversified and will be riskier unless its more heavily-weighted countries (the larger ones!) are much safer than those which it weights lightly.

Equal-weighting requires you to give higher weightings to small markets and, apart from diversifying political risk, it has the additional merit of causing you to emphasise smaller, more 'valuey' markets which are earlier in their development phase. So our intuition is that you can improve the risk and the return characteristics of both emerging markets indices and your portfolios from a simple strategic decision to use equal country weights.

Recently, we decided to test these ideas empirically using the IFC-Investable Index data. We started by reconstructing the capitalisation-weighted index to account for the transaction costs that are really associated with maintaining that index. So, as companies or countries were added or deleted from the index, we subtracted the estimated transaction costs that would have been incurred.

Doing the same for equally-weighted indices is not quite as simple. There is an extra decision, which is how often the index should be rebalanced to benchmark weights. Because you have to sell those countries which have made relative gains to buy those which have lost, the more often you rebalance, the higher you would expect the cost to be. So we repeated the calculation separately for equal-weighted portfolios which are rebalanced monthly, quarterly, and annually to get an accurate sense of the returns.

The diagram shows what were the possible portfolios for a US-dollar based investor combining a diversifed developed-markets equity portfolio with an emerging markets portfolio in the selected interval. These are the efficient frontiers generated by combining the MSCI World Index with each of our modified series of indices in constant proportions.

The higher returns for similar levels of risk from each of the equal-weighted portfolios suggests that the risk in these markets is disproportionately event-driven, and that it tends to be uncorrelated across markets. There are plenty of examples of political crises and currency crises such as Mexico and southeast Asia. These are country-specific or regional, but rarely global, effects. They are not obviously related to the size of the market either, and it may be that the largest of these markets (including the Mexican and Malaysian examples) face similar event risks to the smallest.

The equal-weighted emerging markets portfolios also provided higher returns than their capitalisation-weighted equivalent, even after taking the higher transaction costs into account. This is consistent with a behavioural-finance theory of inefficient markets. Smaller markets incorporate a greater risk premium, but the risk can be diversified in the final portfolio without reducing your stake in the higher expected returns.

The third interesting result is how much better the equal-weighted portfolio did, in terms of return, when rebalanced more often that annually. Again, this is consistent with the behavioural finance literature on speculative bubbles. These country returns are subject to bubbles, just think of the recent performance of South-East Asia. A rebalancing discipline which forces you to sell winners and buy losers each time you return to equal weights exploits the tendency of bubbles to burst.

However, there was practically no difference in the combinations from monthly and quarterly rebalancing, which means that there is some point at which increased turnover wipes out any advantage to higher-frequency recalibration.

The reality is that event risk and political risk are rarely independent of size. In other words, the smallest countries probably are more vulnerable. But the best way to limit investment risk may still be to limit event risk by modifying strategic weighting schemes.

The key conclusive point is that investors should perhaps think about how to allow their managers to exploit behavioural irregularities more directly and actively. So the possibly uncomfortable choice of an equal-weighted benchmark is bound to blur the distinction between investment patterns and strategic allocation decisions, just at the time when investors are so closely focused on sterile benchmark-index selection.

Colin McLatchie is chief operating officer and chief investment oficer with Panagora in London