A Conservative Tactical Momentum Strategy Using Bond ETFs
Summary A simple tactical asset allocation strategy is described that selects one ETF each month from a basket of 5 essentially non-correlated bond ETFs. 4-month and 2-month returns are used to determine the top-ranked ETF each month. The backtest results of the strategy (1998 – present) using mutual fund proxies for the ETFs indicate a CAGR of 14.8% and a maximum monthly drawdown of -10.7%. There are no negative annual returns even in 2001, 2002 and 2008. Any investor can easily set up and trade the strategy using the free Portfolio Visualizer software. In the current market turmoil, many investors are looking for a more conservative bond strategy to reduce risk. In this article, I present a new bond tactical asset allocation strategy that is relatively simple and may be attractive to such investors. One of the obstacles in developing a tactical bond strategy is the short history of bond ETFs. This allows rather limited backtesting of bond ETFs. In order to extend the backtesting timeframe, mutual funds that have longer histories are selected as proxies to the ETFs. The proxy mutual funds are selected based on correlation and performance with their ETF counterparts. The strategy was developed using the free Portfolio Visualizer (PV) software. I first selected bond assets that were relatively non-correlated to each other. Shown below is the basket of assets I decided to use; the basket is similar to the basket used by Frank Grossmann in his “Bond Rotation ‘Sleep Well’ Strategy.” I have listed the ETFs followed by its mutual fund proxy and the average correlation coefficient between ETF and mutual fund proxy. Convertible Bonds: SPDR Barclays Capital Convertible Bond ETF ( CWB) – I nvesco Convertible Securities Fund ( CNSAX) (Correlation = 0.84) High Yield Bonds: SPDR Barclays Capital High Yield Bond ETF ( JNK) – Fidelity Advisor High Income Advantage Fund ( FAHDX) (Correlation = 0.63) Long Term Treasury: iShares 20+ Year Treasury Bond ETF ( TLT) – Vanguard Long Term Treasury Fund ( VUSTX) (Correlation = 0.98) Short Term Treasury: iShares 1-3 Year Treasury Bond ETF ( SHY) – Vanguard Short Term Treasury Fund ( VFISX) (Correlation = 0.80) Emerging Market Bonds: PowerShares Emerging Markets Sovereign Debt Portfolio ETF ( PCY) – T. Rowe Price Emerging Markets Bond Fund ( PREMX) (Correlation = 0.40) PREMX is the only proxy that has somewhat poor correlation to its ETF cousin PCY, but it was the best I could do. A chart from Stockcharts that presents the performance of PCY and PREMX is shown below. The chart shows that PREMX is a reasonable proxy for PCY. (click to enlarge) Asset correlation coefficients for the mutual funds for 07/28/1997 – 08/18/2015 based on daily returns are shown below. The correlation coefficients from PV are the overall coefficients over the entire backtest period. The mutual funds are generally non-correlated or negatively correlated, although a few of the correlations are greater than what I wanted. (click to enlarge) I selected the top-ranked fund each month based on relative strength returns of four months and two months. The overall rankings were based on a weighting of 51% on the four-month rankings and 49% on the two-month rankings. The results taken from PV in tabular form and graphical form (with permission) are shown below. This strategy has a CAGR of 14.8% and a maximum monthly drawdown of -10.7%. This compares to a CAGR of 6.5% and a maximum monthly drawdown of -51% for the S&P 500. It would have been better to use a bond-based fund instead of the S&P 500 as a benchmark, but that option was not available in PV. (click to enlarge) (click to enlarge) There are no negative return years; the worst year is 2001 with a return of +1.5%. The annual returns each year are shown in the table below. The final step in backtesting this strategy is to compare the results of the ETFs with the results of the mutual funds from 2010 – present. The 2010 start date was dictated by the origin of the youngest ETF in the basket. The results for the ETFs and the mutual funds are shown below. The ETFs give similar results as the mutual funds. ETF results (2010-present): (click to enlarge) (click to enlarge) Mutual fund results (2010-present): (click to enlarge) (click to enlarge) The same strategy using the ETFs was executed with the commercial ETFreplay software. The results are shown below. The ETFreplay results were comparable to the PV results, but the ETFreplay results did not exactly replicate the PV results. The CAGR is 15.8% for ETFreplay and 17.4% for PV from 2010 – present. The maximum daily drawdown in the ETFreplay trading scheme is – 10.9%. The maximum monthly drawdown using PV is -6.1%. (A monthly drawdown calculation will always be better than a daily drawdown calculation.) These differences between ETFreplay and PV are primarily caused by ETFreplay using trading days to calculate returns while PV uses calendar months. Also, the data sources may not be the same. (click to enlarge) Another option with this strategy is to select two funds each month instead of only one fund. As would be expected, the performance is reduced, and the risk is improved (i.e. lower drawdown). In the two-fund scenario from 1998 – present using PV, the CAGR is 12.5% and the maximum monthly drawdown is -7.0%. The two-fund scheme also has positive returns every year. Any investor can go to the PV website and run these calculations and determine the monthly ETF selection at no cost. This momentum strategy seems to have potential as a conservative investment strategy based on backtesting to 1998, with the usual caveat that the strategy depends on how faithful the mutual fund proxies mirror their ETF counterparts. Please use this strategy at your own risk. Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. (More…) I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.