When investing in shares, it is often said that time in the market counts and not the timing. However, adopting this strategy can severely hamper returns if you are not careful. By way of example, we will backtest two opposing investment strategies conducted over the last ten years with our flagship PowerShares-A methodology.

The first method will be annual rebalancing, where you create your portfolio, sit back for 12 months and then liquidate the portfolio and start over again. It would appear this strategy has the odds stacked against it as you have the brokerage fees for the buying and selling to consider and you will incur income taxes on the shares you sell as you have not kept them for at least 3 years (SARS give you a "safe harbour" of 3 years after which you only incur 10% capital gains taxes).

The second method is 3-year rebalancing, where you create your portfolio, sit back for 36 months and then liquidate the portfolio and start over again. This method seems to have everything going for it - you incur far less brokerage fees, and you only have to pay SARS capital gains taxes (effectively 10% versus 40% marginal tax in the previous method).

Both strategies above will have portfolios that accrue dividends (remember PowerShares-A are designed to be aggressive dividend yielders), but under current tax laws, these are tax-free in your hands. We will show that annual rebalancing and 3-year rebalancing deliver more or less the same returns depending on your time frame and market cycle.

EFFECTS OF PERSONAL TAXES

In our detailed introduction to the PowerShares-A investment methodology, we never covered taxes. We will now show how personal taxes at the marginal 40% tax rate will have effected your returns over the 10 year period from 1999 in an annually re-balanced portfolio, with dividends re-invested.

We see that taxes have quite a debilitating effect on your returns. Before you throw up your hands, consider that 28% CAGR NET returns after all taxes is still an awesome strategy that significantly out-performs the ALSI and most professionally managed funds. You are more than doubling your money every 3 years on average, AFTER TAX. Also, we believe there is a loophole in the tax laws, which PowerShares can exploit by their nature, but we are busy with legal advice on this issue and will report in due course. For now we are perfectly happy with doubling our money every 3 years.

The first thing that comes to mind then, is "Why not rebalance every 36 months to avoid taxes and brokerage charges?". Let's investigate.

COMPARATIVE TESTS

We will now attempt to compare the two different strategies of rebalancing 12 monthly versus 36 monthly over the last decade. In all tests we include the costs of brokerage charges, income taxes and capital gains taxes as applicable to the various strategies. Dividends are included in the returns calculations.

3-YEAR TEST PERIODS

In our first tests, we will examine 8 sets of 3-year investment periods, namely 1999-2002, 2000-2003, 2001-2004,....,2006-2009. The 12 month rebalancing strategy will have to rebalance twice in each investment period and pay 40% marginal taxes on profits of all share proceeds before re-investing, whereas the 36 month rebalancing strategy only incurs 10% capital gains tax on the share proceeds at close-out of the investment period.

The 12 month strategy will re-invest dividends accrued in the rebalancing exercise, whereas the 36 month strategy will merely accrue dividends as cash during the 3 year investment period. The 3 year rebalance strategy cannot really re-invest the dividends after years 1 and 2 since they never get big enough to re-invest without crippling brokerage charges, unless your portfolio size is above R500,000-R800,000 (unless an individual share purchase is above R5,000, brokerage as a percentage of the transaction is greater than 2%. This implies you need 5 x R5,000 = R25,000 in dividends to re-invest at a minimum in all your shares in the portfolio. At an average yield of 5% this implies portfolios size greater than R500,000) . The 12 month strategy does not have this problem since it is at critical mass at each re-investment period due to the rebalance.

The cumulative after-tax returns are shown below for each of the 8 test periods. We note that the 1-year rebalance strategy had a win-rate of 3 out of 8 (37.5%) versus the 3-year rebalance strategies 5 out of 8 win rate (62.5%). However when averaging all the cumulative returns over all 8 periods, the 1-year strategy at 111.3% is marginally less than the 3-year strategies' 120.4% More importantly though, the standard deviation of test-period returns is FAR LESS for the 1-year balancing strategy (30%) versus the 3-year balancing strategy (65%). This tells us that from a risk-adjusted perspective, the 1-year strategy is the less volatile option.

It is comforting to see that regardless of which strategy you chose, ALL the 3-year test periods over the 10 years comfortably doubled your initial investment (more than 100% cumulative return). Note how the 3-year strategy succumbed to the 2008/9 bear market plunge, versus the 1-year strategy which carried on as if the bear market never existed.

6-YEAR TEST PERIODS

We will repeat the above exercise, but with longer investment periods. We will examine 5 sets of 6-year investment periods, namely 1999-2005, 2000-2006, 2001-2007,2002-2008 and 2003-2009. The 12 month rebalancing strategy will have to rebalance 5 times in each investment period and pay 40% marginal taxes on profits of all share proceeds before re-investing, whereas the 36 month rebalancing strategy will only have to rebalance once and incur 10% capital gains tax on the share proceeds before re-investing.

As before, the 12 month strategy will re-invest dividends accrued in the rebalancing exercise, whereas the 36 month strategy will merely accrue dividends as cash except for when it rebalances once after the first 3 years.

The cumulative after-tax returns are shown for each of the 5 test periods. We see that the 1 year strategy achieves a 2 out of 5 (40%) win rate, not too dissimilar to the previous exercise. We note again that when the returns are averaged for the 5 test periods, the difference is marginal whilst the standard deviation is significantly different. Once again from a risk adjusted (volatility) perspective the annual rebalancing seems the wiser option.

9-YEAR TEST PERIODS

We will examine 2 sets of 9-year investment horizons, namely 1999-2008 and 2000-2009. The 12 month rebalancing strategy will have to rebalance 8 times in each investment period and pay 40% marginal taxes on profits of all share proceeds before re-investing, whereas the 36 month rebalancing strategy will only have to rebalance twice and incur 10% capital gains tax on the share proceeds. As before, the 12 month strategy will re-invest dividends accrued in the rebalancing exercise, whereas the 36 month strategy will merely accrue dividends as cash except for its two rebalance occasions.

We notice that the volatility inherent in the standard deviation increases with investment horizon. The 3 year strategy averaged out at 926% cumulative return (29.5% CAGR) for the 9 years and the 1-year strategy averaged out at 738% cumulative return (26.6% CAGR). The 1-year strategy was far more consistent during both 9 year test periods though, with a negligible 58% standard deviation, versus the 3 year strategies' whopping 464% deviation. Again, from a risk adjusted perspective, we seem better of with the annual rebalanced strategy.

CONCLUSIONS

The big surprise was that the annual rebalance strategy managed to hold its own against the much more tax efficient 3-year rebalance strategy. Whilst it would appear that the 3-year rebalance strategy has the higher win-rates and marginally higher average returns, it has over double the standard deviation of returns of the 1-year rebalance strategy, hence far higher risk. If one takes the average returns and divides it by the standard deviation, the 1-year rebalance strategy works out at double the risk adjusted return ratio than the 3-year rebalance strategy.

Our observation is that the 1-year rebalance strategy perfectly suits the profile of the PowerShares portfolios, namely, good quality, large, productive companies going for cheap that realise their value within a 12-18 month time frame. After the 12-18 month period during which the share "recovers" to fair price, then it is subject to the vagaries of the market and the investing public and this is when things get less certain and growth gets slower. PowerShares portfolios older than 18 months in general start adopting a closer correlation to the general market and can often become overpriced, exposing you to more risk. All you need is one disappointing earnings result or a market correction and the share plummets.

By contrast, younger PowerShares are by definition under-priced and have lower correlation to the general market. They quicky rise back to fair value or higher than fair value within 12 months and are aggressive dividend yielders (due to the low price paid for them) and it is this characteristic alone that allows the gains of the 1-year rebalance strategy, even with higher taxes and brokerage charges, to meet up with those of the tax efficient 3-year rebalance strategy. It is also this characteristic that contributes to the superior risk adjusted returns brought about by the low standard deviation of return.

Our advice is to maintain tax efficient 3-year rebalance strategies on portfolios created at or soon after a bear market trough (such as now), but to switch to 1-year rebalance strategies the minute any correction greater than 20% occurs in the general market or your portfolio. When the market correction appears over, switch back to 3 year rebalancing. If you feel the market is overheated and is likely to suffer a correction, stick to 1-year rebalancing. As a general rule of thumb, the minute a PowerShares-A portfolio shows signs of slowing growth or decline after the initial 12 months growth period, close it out and rebalance.

If the dividends yielded from your portfolio are in excess of R25,000 per annum (ie you are running large portfolios) then 3-year rebalancing strategies with re-invested dividends are likely to always outperform the annual rebalance strategy (after taxes), but you will be subjected to greater volatility in annual returns. However, our tests show you would be far better off taking those dividends and using them to create a brand new PowerShares portfolio (consider this as partial annual rebalancing!) to run in parallel with the existing portfolio that generated the dividends.