Systematically rebalancing portfolios can increase returns and reduce risk for clients. Yet it is a process practiced by few financial planners.
By rebalancing, I mean the periodic assessment of asset allocation against a desired model and, if the divergence is too great, making appropriate adjustments. How often this should be done, and what margin of divergence is appropriate before acting, are issues dealt with below.
It must be acknowledged that there is ongoing work in rebalancing. ‘Auto rebalancing’ platforms are available, which are better than nothing, but these lack tailoring and can, for example, incur otherwise avoidable capital gains taxes. Careful rebalancing would assess whether to make adjustment within or outside superannuation, which particular fund to realise gains from, the timing of this, and so on.
I’ve heard claims that rebalancing adds no value. As with many false industry clichés, this statement can and should be subjected to testing. There are market conditions in which it has little effect — for example, where markets run broadly in parallel. However, in this circumstance, little systematic rebalancing is triggered.
There are also cases where the rebalancing process detracts value for a time, for example, where an asset class is consistently outstanding or dreadful for a number of years, but scenarios like this are in a clear minority.
Let’s start with an intuitive analysis before beginning statistical verification.
The difficulty in timing markets is widely acknowledged. Yet the danger of never taking profits from an asset class is equally well recognised, at least since the bursting of the technology bubble. One advantage of regular rebalancing is that profits are systematically extracted at relatively high points in markets. For the asset to have approached a peak, it must have performed strongly, hence growing as a proportion of the portfolio, and consequently being reduced.
Equally, rebalancing increases holdings of assets at a low point, via the same process. Is buying low and selling high desirable?
Table 1 shows the return and volatility of four assets since 1980 in descending order of return: international shares, listed property, domestic shares and bonds.
Table 2 shows the results achieved by three mixed portfolios: shares/bonds; international shares/domestic shares, and international shares/property. In the first two cases, we diversified 25 per cent out of the superior returning assets and introduced an inferior one. With international shares and property, the two top performers, we diversified 50/50. These are rebalanced annually at year end.
In every case, the diversified and rebalanced portfolio equalled or beat the return of the better asset class and reduced its volatility.
So much for the argument that rebalancing adds no value. It implements a progressive realisation of profits at higher points in a cycle accompanied with a progressive increase in exposure at lower points in a cycle. It would be fairly extraordinary if this did not provide a satisfactory outcome. Yet it is carried out through a discipline that requires no market timing.
As an aside, these examples are further counter evidence to the ‘long term equities always win’ mantra. Even if the profit from shares has exceeded bonds since 1980, adding 25 per cent bonds is better again. International shares have been the top class, but adding 50 per cent property produces an additional 0.4 per cent per annum.
It is important to explain that these results would not be achieved without rebalancing. Without this, the return is always between that of the two assets individually — as is logical. Only by realignment is the extra profit realised.
If planners don’t rebalance, they are sacrificing one of the key benefits of diversification and giving up return.
Another positive for a portfolio adjustment discipline is that it ensures clients’ assets reflect current, rather than historical, asset allocation models. There is a school of thought that ‘balanced’ portfolio mixes never change — being heavily equity biased at all times. This is ridiculous. Anyone who held a constant view on asset allocation over the period in the above examples must not have noticed the disappearance of inflation.
Bond yields in the mid teens produced a negative net real return for high taxpayers in the 1980s. At around 6 per cent with inflation around 2 per cent, they are now positive after tax. Asset allocation should recognise this (as should those commentators who continue to refer to today’s ‘low rates’).
Asset mixes now may also include ‘alternatives’ (for example, hybrids or hedge funds) and presently favoured fund managers. Rebalancing to a current model ensures existing clients remain up to date.
One cost of rebalancing can be capital gains tax, reducing the pool of capital for compounding. This is especially a problem with auto-rebalancing platforms. Frequency of realignments is another factor in this.
Clearly, if charges are within allocated pensions or superannuation, tax is irrelevant or minimal respectively. Further, with portfolios that are either added to or drawn from, rebalances are made by additions or withdrawals, not sales.
In other cases, tax is a factor that does need to be taken into account. Yet, ironically, rebalancing is most profitable where tax liabilities can be highest. Who would argue that rebalancing out of international equities at the peak of the tech boom was not worth the tax (especially as rebalancing also guaranteed assets were re-added at the low points in 2003)?
When should realignment occur? With an asset that periodically distributes realised capital gain, like many international share funds, post-distribution dates is far preferable as a rebalance date.
Further, our modelling demonstrates it should not be done daily, nor on every occasion there is a small divergence (which would also maximise capital gains tax). In practice, once or twice a year produces favourable outcomes. If a planner’s system allows, whenever divergence reaches set limits is better.
When divergence of an asset class from the desired exposure is not material, no change is justified. Modelling suggests that 10 per cent should be a minimum trigger (that is, if an asset should represent 25 per cent, a variation, plus or minus, of 2.5 per cent calls for attention).
A wider margin of up to 20 per cent (that is, plus or minus 5 per cent against a 25 per cent target) also adds returns, and reduces planner workload.
It may add strength to the argument to take an extreme example over a longer period. Since June 30, 1969, a portfolio of 50 per cent shares and 50 per cent cash (surely a stupid asset mix) outperformed 100 per cent shares with half the volatility — if rebalanced whenever divergence was 15 per cent.
If putting half one’s capital in cash for almost a quarter of a century adds return, rebalancing must be fairly productive.
Systematic rebalancing adds returns, reduces volatility and ensures planners don’t leave clients with maximum exposure at market crests. Just as important, it is seen to add value at difficult times and in periods where prospective returns are low. Why then, do not all planners practice it?
Perhaps because it demands ongoing work. But this is what we are paid for — which of course assumes we are paid for ongoing service.
Robert Keavney is chief executive officer of Centrestone Wealth Advisory.
