If you are new to the concept of rebalancing, we suggest that you read through our earlier articles "As The Year Comes To A Close, Consider Rebalancing" and "Mythbusting: Frequent Rebalancing Improves Portfolio Return" before proceeding.

Questions Asked:

• How do sales charges and switching fees affect the returns of the example unit trust portfolio shown, i.e. the one that is rebalanced annually?
• Does Dollar Cost Averaging (DCA) work well with rebalancing?
• Is it necessary to backtest from as long as 41 years ago? How about shorter periods? Will rebalancing beat buy-and-hold method?
• What if unit trusts are used instead of indices for backtesting?

Though we have provided the answers for the above questions, we believe that other investors may have the same questions so having the answers up on our website would be useful, and we wanted to provide data to support to our answers.

Q1: How do sales charges and switching fees affect the returns of the example unit trust portfolio shown, i.e. the one that is rebalanced annually?

• As presented during the event, we did a mechanical rebalancing backtest from end-1970 to end-2010 to compare the returns of a unit trust portfolio with 100% allocation in equities (breakdown in Table 1). The results of this comparison is shown in Chart 1, where we plot the returns of 1) a portfolio with annual rebalancing, 2) a portfolio using buy-and-hold method, and 3) the MSCI AC World Index which has a different composition than our example portfolio.

  Table 1: Unit Trust Portfolio's Index Allocation
  Country	Indices	Index Allocation
  US	S&P 500 Index	30%
  Europe	MSCI Europe Index	30%
  Japan	Nikkei 225	10%
  Korea	KOSPI Composite	7.5%
  Hong Kong	Hang Seng Index	7.5%
  Singapore	FTSE Straits Times Index	7.5%
  Taiwan	Taiwan Weighted	7.%
  Source: iFAST compilations

  Chart 1: Mechanical Rebalancing Backtest Without Sales Charges and Switching Fees

  
  

  Now for the fun part, we incorporated a 2% sales charge on the RM10,000 initial investment and a 2% switching fee on each rebalancing transaction performed. Chart 2 shows the results.

  Chart 2: Mechanical Rebalancing Backtest With Sales Charges and Swithcing Fees

  With the inclusion of transaction costs, the gap between the returns of the rebalanced and the buy-and-hold portfolio has narrowed, but this is to be expected since the rebalancing method incurs extra switching fees. However, the rebalancing method still outperformed both the buy-and-hold method and the MSCI AC World Index, both of which only incur a 2% sales charge on the RM10,000 initial investment.

Q2: Does Dollar Cost Averaging (DCA) work well with rebalancing?

• The examples we have shown above are based on lump sum investments, but keen investors wanted to know if DCA would improve the returns of a rebalanced portfolio. We are a firm believer of the DCA method when investing in equities but putting our bias aside, we are not expecting the DCA to outperform a lump sum investment because of the simple fact that equities have been on a general uptrend over the last 41-year period.

  For this test, we used the same data as in Chart 2 and incorporated annual purchases of RM10,000 into 1) the rebalanced portfolio, 2) the buy-and-hold portfolio, and 3) the MSCI AC World Index. Sales charges and switching fees of 2% were applied on each transaction, and the results are shown in Chart 3. One thing to note is that the total investment amount in Chart 3 at the end of the 41-year period is RM420,000 (RM10,000 x 42), instead of only RM10,000 in Chart 1 and Chart 2.

  Chart 3: Mechanical Rebalancing Backtest With Dollar Cost Averaging (DCA)

  Surprisingly, the rebalancing method with DCA managed to outperform the rebalancing method without DCA, albeit by only 23.85% points over the 41-year period. Looking at the buy and hold method, the lump sum investment outperformed DCA by a margin of 414.25% points.

  As such, the strength of DCA lies not in its ability to outperform lump sum investments – in this case it clearly doesn’t, but in its ability to deliver robust returns without having to come up with all the capital at the outset. Our earlier article explains the pros and cons of the DCA method.

  Using DCA and rebalancing your portfolio gets a little messy and investors may use softwares like Microsoft Excel to put their investments in order. Some investors may also argue that a difference of just 23.85% points over a 41-year period is too little considering the extra effort needed, so your mileage may vary.

Q3: Is it necessary to backtest from as long as 41 years ago? How about shorter periods? Will rebalancing beat buy-and-hold strategy?

• To find out, we narrowed our test window down to 5-year rolling periods, as this represents a long-term investment horizon that we are confident in. Based on annual data points, we took 5-calendar year periods (from end of 1970 to end of 2010) and assessed the return of each period (there are 37 periods tested).

  We then arranged the results to show the probability of achieving a certain range of returns (see Chart 4).

  Chart 4: Probability of Returns

  Based on Chart 4, the rebalancing method has a lower probability of making losses as compared to the buy-and-hold method (refer to the -30% to 0% range; the largest loss was 30.0% for the 5-year period ended 2003). Conversely, this means that if you rebalance your portfolio, the probability of achieving a positive return is higher than if you bought-and-held. Apart from that, the rebalancing method had a surprisingly better record in the highest-end of the range - over 120% returns. What this means is that cashing out of outperforming markets does not limit your gains, despite the common tendency for an investor to hold his or her top performing investments in the belief that they will chart new highs for at least the 'near future'.

Q4: What if unit trusts are used instead of indices for backtesting?

• The purpose of using benchmark indices to perform our tests is because most unit trusts on our platform do not have a long enough track record. As with any statistical test, the larger the population and the more data sampled, the more meaningful the results. Also, a longer period is needed to capture different market cycles in order to get a clearer picture on the performance of the different investment methods.

  Moreover, we believe that a unit trust will most likely perform in tandem with the particular market it invests in. However, there is a also the possibility that the unit trust does not perform although the market it invests in is performing. 

  Our advice to investors is to review their portfolios at least annually to assess the performance of their unit trusts compared to their respective benchmark indices and peers. Investors can also leverage on our Recommended Funds List to shortlist the funds they want to invest in.

Conclusion

We hope this article is able to clear some of your doubts on rebalancing. If you have further questions on rebalancing, please do not hesitate to email us. We will try our best to answer your questions.

Related Articles:

Mythbusting: Frequent Rebalancing Improves Portfolio Return
Idea Of The Week: 5 Tips To Rebalance Like A Pro
Dollar Cost Averaging Method
Dollar Cost Averaging With A Regular Savings Plan
Tested: Lump Sum Investing vs Dollar Cost Averaging
Key Investment Themes And 2012 Outlook