Recently I began studying the question of how to maintain one's portfolio for maximum risk avoidance and profitability. My own preference is to use ETFs, both long and inverse styles, so my research is based on these. However, I can see that this technique will adapt in a similar fashion to stock trading with little variation.
I developed several systems using ETFs that give signals when to buy or sell. When my system indicates a long position, I would buy the long ETF of the underlying index and to sell I would buy the inverse ETF. For this study, I only paper traded the S&P 500 index, the T-Bond index, and the VIX index. For proxies I used the Direxion SHS ETF Daily Bear fund (NYSEARCA:SPXS), the Direxion SHS ETF TR DRX S&P500 Bull fund (NYSEARCA:SPXL), the Credit Suisse Nassau BRH VIX Inverse Fund (NASDAQ:XIV), the Barclay's Bk Plc S&P500 VIX fund (NYSEARCA:VXX), the Direxion SHS ETF TR 20yr Treasury Bull Fund (NYSEARCA:TMF), and the Direxion SHS ETF TR 20yr Treasury Bear Fund (NYSEARCA:TMV).
I chose these because the data for them are available, and I am testing the idea rather than the final results per se. My systems give buy/sell signals from an optimized MACD I developed some time ago and is always in the market, either long or short. I use Excel to develop my systems, but coordinate with the charts (various time frames) to do my actual trading. My systems use the Friday close each week.
This particular system of adaptation contains two elements: When should the portfolio be readjusted to the total portfolio proportions? Should the proportions be equal or adjusted to the original proportions? Many systems merely set the proportions in the beginning and look at each investment as an individual element. I wanted to see if re-proportioning a portfolio outperformed that system or not. Since each signal is optimized, I didn't look at that aspect.
The historical profit of each system and the Sharpe index for each acted as guidelines for the original proportions. I then played with the numbers to find the optimum profit/Sharpe combination. The Sharpe is just a measure of the profitability of a portfolio when compared to the market return. The higher the Sharpe ratio, the better. My testing indicated that I should use 50% S&P, 10% T-Bonds, and 40% VIX. Some proportions gave higher returns but lower Sharpe indexes, so these ratios were subjectively chosen.
The chart showed low volatility and consistent gains, though there are occasional dips. I don't think that short-term losses are avoidable because any mechanical system will show these. The long-term results are what one seeks to optimize.
Next, I wanted to compare various re-proportioned timings. First, do re-adjusted proportions help the final result? If so, how often should it be done? The answers to these questions are not intuitively obvious. Proportioning involves what percentages to allocate to each vehicle, equal or unequal. Then, should each component be reallocated to the original proportions or should they be given equal parts of the previous week's total?
The following chart shows the results of my study:
In the chart above, which I created in Excel, the blue line is the value of the return in the original version (total) without re-proportioning. Each vehicle used only the profits it has earned to reverse. The red line is the return from equal proportions and re-apportioned monthly. The green line used equal proportions but reapportioned weekly (one-third of the total portfolio to each vehicle).
Finally, the orange line is the return from a weekly reapportioned sum (using the original proportion for each). The final return for "Total" was 176% per year average, for "Rebalanced System 3- all 3 weekly" was 163% per year average, for "Rebalance System 1 monthly" was 71.6% average annual return, and "Rebalance System 4 All Weekly" was 182% average annual return. The Sharpe ratio for "Rebalance System 4 All Weekly" was 5.08. These are remarkable numbers all by themselves, but the improvement from "Total" to "Rebalance System 4" is considerable.
It turns out that one should re-apportion to the original allocation each week (using the Friday close). This may vary some for your particular set of stocks or ETFs. The return shows a much higher Sharpe index, lower volatility, and greater final return. All the reallocated funds had lower volatility than the "Total" system. The time frame was from November 2009 through June 2014. Data was the weekly close downloaded from Yahoo Finance.
When using this system of portfolio re-apportionment, the total portfolio size must be taken into account. For portfolios less than $10K, the transaction cost may absorb the extra profit. I also see no reason to limit the number of ETFs or stocks for this system. I used three vehicles, but up to six seem to present no problem. I hesitate to use more than six in real trading because following more may become too difficult a task. The money management rules that I follow allow no more than six, no less than three at a time.
To use the system with stocks, I would go long or short stocks that one is familiar with, probably with higher volatility. The other problem with stock is that there are many other variables to consider, requiring far more time and effort. To use other criteria and only go long would require doing a study based on one's own primary system. Using ETFs reduces this problem significantly. I did not need to study a system using a hundred stocks but only 18 ETFs, in nine markets.
In one set of trials I did, I used the Utility Index (using UPW and SDP). This process reduced volatility significantly, but it reduced the final return so much I decided not to use it. Determining which ETFs to use is an individual choice, and is based on one's tolerance for risk. The general rules I use to determine which ETFs to follow are that they should be multiple funds (2x or 3x, positive and inverse) because of the higher return and high volume to increase liquidity when revision is necessary. Higher multiple ETFs also reflect higher volatility, making signals a little easier to recognize.
By utilizing some consistent buy/sell system reduces this volatility, and using several systems reduces the portfolio risk. Some markets are highly profitable, but generating signals or selling the fund may be so difficult that the profit is lost. Selling short the ETFs (to create synthetic long or short positions using the convexity property of multiple funds) provides far greater profits, but it is nearly impossible to short them.
This study demonstrates that rebalancing improves results both in consistent profits and reduced volatility, improving the Sharpe ratio of a portfolio. By rebalancing each week my total portfolio to the original proportions of my ETF systems, caused my return to grow, lower volatility, and increased the consistency. In my opinion, every investor could benefit by rebalancing, though every portfolio would require individual analysis to determine percentages allocate to each part.
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.
Disclaimer: My results are based on my personal systems and money management techniques. Your results will vary depending on your own systems.