Rebalancing is the act of bringing one’s investment portfolio to its “original” state of predetermined asset class allocation levels, and it is one of the core tenets of diversification. Typically for passive investors, rebalancing is done once or twice a year; this frequency is seen as balancing between maintaining the portfolio and the transaction costs that come along with rebalancing.
With the advent of brokerages offering less expensive commission charges and fees, the costs of rebalancing monthly or quarterly may be equal or below to that of semi-annual or annual rebalancing using brokerages that still charge commissions at prices a decade ago. This is highly plausible if a passive investor only has 5 or less securities (especially exchange traded funds, or ETFs) in his/her portfolio.
On this thought, I decided to carry out a little test to see if the rebalancing frequency affects returns. Using the data from Portfolio Visualizer1 (www.portfoliovisualizer.com), I used three portfolios: the traditional 60/40 equities/bond, the Bogleheads Three-Fund (50% United States (US) equities, 30% global ex-US equities and 20% bonds) and the balanced Bedokian Portfolio (35% equities, 35% REITs, 20% bonds, 5% commodities using gold and 5% cash).
Three rebalancing scenarios are used: monthly, quarterly and annually. The initial amount would be US dollars (US$) 12,000, and subsequent injection would be US$ 1,000 per month for monthly, US$ 3,000 per quarter for quarterly and US$ 12,000 per year for annually. The period tested would be from January 1994 to December 2021 (this period is used as the earliest REIT data obtained is from January 1994).
|Period Jan 94 - Dec 21||Portfolio Returns (US$)|
|Rebalancing Frequency||60/40 Equities/Bond||Bogleheads Three Funds||The Bedokian Portfolio|
Fig. 1: Portfolio returns with monthly, quarterly and annual rebalancing, Jan 1994 to Dec 2021. Figures are before inflation.
Interestingly, returns using quarterly rebalancing came out tops, with monthly rebalancing coming in second and annual rebalancing third.
Let us look at another set of results, the average 10-year and 15-year rolling returns. These two timeframes are used on the assumption that an investment portfolio is maintained for at least 10 years.
|Period Jan 94 - Dec 21||Portfolio|
|Average Rolling Returns (Rebelancing Frequency)||60/40 Equities/Bond||Bogleheads Three Funds||The Bedokian Portfolio|
|10 Years (Monthly)||6.47%||6.78%||8.48%|
|15 Years (Monthly)||6.26%||6.44%||8.08%|
|10 Years (Quarterly)||6.59%||6.88%||8.55%|
|15 Years (Quarterly)||6.38%||6.54%||8.16%|
|10 Years (Annually)||6.63%||6.96%||8.51%|
|15 Years (Annually)||6.44%||6.63%||8.10%|
Fig. 2: Average 10-year and 15-year rolling returns with monthly, quarterly and annual rebalancing, Jan 1994 to Dec 2021.
Surprisingly, the results were mixed. Annual rebalancing brings about a higher average rolling returns for the 60/40 equities/bond and Bogleheads Three-Fund for 10-year and 15-year, while the Bedokian Portfolio’s quarterly average rolling returns were better than the other two variations.
Do note that the data used is based on the perspective of the US markets, and transaction costs and taxes are not factored in.
Explanation And Conclusion
From a prima facie observation, the dataset used for the asset classes were the same throughout the three portfolios, hence there is a probability that the individual asset class(es) themselves are displaying better returns by quarterly capital injection. There is also the possibility that for quarterly rebalancing, the securities were bought in at lower price points as compared to monthly or annually, where the latter two could be bought in at higher price points.
As for the average rolling returns, since it is an average reading, the actual numbers that form the average may see more deviations around the mean, which may explain why the annual average 10-year and 15-year rolling returns were higher even with quarterly rebalancing bringing in more returns (for the 60/40’s and Bogleheads’).
While looking through the figures of different timeframes (at least 10 years) of different portfolios using Portfolio Visualizer (which I had not published here), it seems that returns from quarterly rebalancing were relatively better than those from monthly and annual ones. Perhaps going quarterly is better?
Then again, we cannot have this conclusion due to two simple reasons: first, I had only utilized one data source. The dataset used is based on the presumptions that Portfolio Visualizer had stated, such as the sources and method of calculations. It is recommended that other sources (and their underlying basis of data) are to be used so that we could have a fair comparison across.
Second, and more importantly, all these are inferences and results from past data, and we all know that past performances are not indicative of future results.
Stay safe and stay vested.
1 – https://www.portfoliovisualizer.com/backtest-asset-class-allocation