Restrict rebalancing to US stocks and bonds: Morgan Stanley

A more efficient way to rebalance highly diversified multi-asset portfolios – which contain illiquid assets – could be to restrict the rebalancing to exchanges between US stocks and US bonds only, according to new analysis by Morgan Stanley.

The gain or loss from high frequency rebalancing has been shown to be surprisingly small, according to Morgan Stanley research, now that premise has been extended to the effects of rebalancing in a more diversified multi-asset portfolio, that specifically holds more illiquid assets.

In previous research Martin Leibowitz and Anthony Bova, found that lower frequency or beta-target rebalancing was shown to have a significant advantage, in the context of a simple 60/40 portfolio, in terms of asset value and transaction volumes, but with a disadvantage in the form of increased tracking error.

Leibowitz is managing director of Morgan Stanley’s US equity strategy team, and Bova is a vice-president of equity research in global strategy.

Their paper explores how these different rebalancing strategies fare when applied to more complex multi-asset portfolios over the 20-year period from 1990 to 2010.

The paper compares annual versus monthly rebalancing back to the initial allocation percentages and finds there was surprisingly little difference in asset values, although the annual approach incurred lower transaction volumes. Both, however, led to sizable drifts in the fund’s beta values.

The paper suggest an alternative rebalancing strategy back to a target beta –  using only exchanges between the highly liquid US stocks and US bonds – allows for more efficient transactions.

The initial portfolio allocations used were US equity 30 per cent, US bonds 25 per cent, international equity 25 per cent, emerging market equities 10 per cent, and US REITS 10 per cent.

According to the paper, when institutional investors set portfolios there tends to be a beta contribution balance between the US equities and bonds sub-portfolio, and the other assets.

For example, with 50 per cent of the portfolio in US equities and bonds, the contribution to overall beta will be in the order of 0.3. The remaining 50 per cent of the portfolio will also have an average beta around 0.6 and so have a similar 0.3 beta contribution.

The analysis shows that the US equities/bonds beta contribution was more stable, before 2005, while the other assets’ beta contribution was between 0.2 and 0.4.

After 2005, the other asset beta contribution has ranged between 0.4 and 0.6, and the US equities and bonds contribution was stable. So the higher betas from the other assets were the main driver of increased overall portfolio beta in recent years.

As the individual asset components move with fluctuations in the equity market, the overall portfolio beta can deviate significantly from the target level.

One possible rebalancing approach would be to reset the portfolio beta to 0.65 at the end of each month. By restricting the rebalancing to US equity and bonds, a more cost-effective approach can be pursued.

In the analysis, the authors use an example of beta target rebalancing if the equity market falls by 30 per cent.

Given the portfolio asset weights (above) initially international equities, emerging markets and real estate contributed 0.31 to the overall portfolio beta, while US equities and bonds contributed 0.34.

After a market decline of 30 per cent the new weights and betas for international equities, emerging markets and real estate have increased their total portfolio beta contribution to 0.36. So the authors say, an additional 0.29 beta contribution must be formed from the combination of US equities and bonds in order to reach the 0.65 target.

This can be achieved by moving 3 per cent from US equities to US bonds.