Asset allocation - is there a hole in the bucket strategy?

Tim Sanderson puts the traditional asset allocation approach head to head with the bucket strategy, with some surprising results.

When advising a client to commence an account based pension, many financial advisers recommend that a ‘bucket strategy’ is used as part of the client’s asset allocation process for risk management purposes.

Theoretically, a bucket strategy that successfully prevents growth assets being sold at temporarily low prices should provide a better outcome than a traditional asset allocation approach over the long- term.

However, little analysis has been done to see whether this approach is likely to work in practice and the effect of differing market conditions and different approaches to implementing and managing the strategy.

How does the bucket strategy work?

Essentially, the bucket strategy involves segmenting a client’s pension balance into two or more pools or buckets:

• A ‘cash bucket’ with safe investments for pension payments expected over three to five years; and
• An ‘investment bucket’ invested in accordance with the client’s risk profile for longer term growth.

The cash bucket is then replenished periodically, with enough funds to cover the next three to five years of expected pension payments. Depending on the strategy approach taken, this could occur:

• At the end of the three to five year period, or earlier if the cash bucket is depleted sooner;
• Each year when the client’s situation is reviewed; or
• Tactically, where the investment bucket cumulative returns have exceeded certain benchmarks, or the investment bucket appears overvalued.

Another important consideration is whether income generated by the investment bucket will be immediately allocated to the cash bucket (minimising the need to replenish this bucket) or reinvested (which may be the only practical option available in non-wrap style superannuation funds).

Historical returns have shown that growth assets such as shares and property are very likely to outperform purely cash based investments over the long-term (the sort of timeframe that normally applies to an account based pension).

In the short-term, however, the stock market can be extremely volatile, as anyone who has endured the global financial crisis would know.

The logic of the bucket strategy is that regardless of the short-term performance of the investment pool, pension payments can continue from the cash pool without the forced sale of growth assets at unfavourable prices.

What is often not discussed though, is that a traditional asset allocation approach automatically benefits from a reverse version of dollar cost averaging – a strategy often used to minimise risk when investing by buying parts of the overall investment at a range of different times and prices.

Assuming pension payments are received on a regular basis (eg, monthly), the sale price of investment bucket assets on a particular day is of reduced importance because many different prices are used throughout a financial year (and over many years).

Analysis

With this in mind, we have compared a traditional asset allocation approach against a bucket strategy approach using historical returns from October 1998 to September 20111.

Our comparison involves a client (age 65) who is commencing a $500,000 account based pension. They will draw pension payments of $25,000 per annum (paid monthly and indexed at 3 per cent annually – always exceeding the required minimum payment).

Under the bucket strategy approach, three years worth of pension payments were allocated initially to a cash pool, which was then run down over three years and replenished at the end of the three year period (with the next three years worth of pension payments).

We can see that, although the account balances vary significantly throughout the timeframe, the traditional asset allocation approach provided for a higher overall pension balance at all times.

By the end of September 2011, the difference in account balance between the two was $6,304, with balances of $551,020 (traditional approach) and $544,716 (bucket strategy approach).

Interestingly though, the difference between the two approaches reached over $35,000 around age 73 and then reduced to $4,200 around two years later, indicating that short-term market movements affect these strategies in contrasting ways.

We previously mentioned a variation on the bucket strategy involving tactically replenishing the cash bucket where investment bucket returns exceeded certain benchmarks.

To demonstrate how this would fare in this example, we assumed that under the bucket strategy approach, the cash bucket would be replenished immediately (with three years worth of expected pension payments) once the investment bucket had returned 10 per cent (including income and capital growth).

A further replenishment would occur with each subsequent 10 per cent return. When compared with the traditional asset allocation approach, results were as follows:

Figure 2 shows that the tactical bucket strategy was less successful than either the original bucket strategy ($9,810) or the traditional asset allocation approach ($16,113) over the period of historical earnings.

While once again the difference varied throughout the timeframe (determined by periods of good or poor market performance), it is clear that the tactical approach risked replenishing the cash bucket more frequently than was really required (in this case, it led to a sustained over-allocation to cash assets).

Rising and falling markets

It is apparent from Figure 2 that the bucket strategy approach tended to perform relatively better during periods of poor market performance (eg, age 74 to 76) and relatively worse during times when markets were performing strongly (eg, age 70 to 73).

The simple reason for this is that, by design, the bucket strategy involved a significantly higher allocation to cash assets, which generally performed better over the short-term than a falling share market.

For example, using the above client, we analysed the same pension over two six-year periods². In the first six-year period (October 2005 to September 2011), a cash portfolio returned 5.4 per cent per annum, while a balanced portfolio returned 2.1 per cent per annum.

Unsurprisingly, the bucket strategy approach outperformed the traditional approach by $4,246.

In contrast, during a second six-year period (November 2001 to October 2007), a cash portfolio returned 5.2 per cent per annum, while a balanced portfolio returned 7.4 per annum.

During this time, the traditional approach outperformed the bucket strategy approach by $5,183.

While analysing short-term market trends assists us to understand the way each strategy works, it is most important to focus on the overall return profile a client can expect over the life of their pension.

History suggests that over many years, an appropriately diversified portfolio invested in line with a client’s risk profile is not enhanced by an artificially high allocation to cash assets.

Conclusion

Based on the above analysis, we have summarised some of the shortcomings of the bucket strategy and why, as a long-term strategy, it doesn’t compare with a traditional asset allocation approach:

• Overweight in cash investments – the bucket strategy approach involves always holding a larger portion of the portfolio (than dictated by the client’s risk profile) in cash assets, which are likely to underperform growth assets over the long-term.
  During short-term market downturns, the bucket strategy approach performs relatively well (due to the higher allocation to cash), but if the long-term trend in markets is up, the long-term trend in the bucket strategy approach will be to underperform the traditional asset allocation approach.
• A successful bucket strategy involves timing the market successfully – the bucket strategy would likely provide a better long-term outcome than the traditional asset allocation approach if the cash bucket was only refreshed (by selling assets in the investment pool) at or near the top of the market cycle, and never when markets are at temporarily low levels.
  While it might be possible to successfully time this refresh some of the time, it is almost impossible to ensure that this is always the case throughout the life of the pension.
• No dollar cost averaging – the traditional asset allocation approach naturally involves dollar cost averaging out of growth assets as regular pension payments are made, reducing some of the risk of short-term market volatility.

Under the bucket strategy, however, a relatively large amount (a number of years worth of future pension payments) needs to be ‘switched’ out of the growth portfolio periodically as a lump sum, which (with the benefit of hindsight) may or may not take place at a favourable price.

Tim Sanderson is senior technical manager at Colonial First State.

¹ Monthly historical returns for Colonial First State Cash Fund and Colonial First State Balanced Fund.

² Six year periods were chosen to allow for two full cash-bucket replenishing cycles.