Over the past few year, I have written a number of articles on Seeking Alpha featuring conservative, low risk, tactical asset allocation strategies that can be used by the do-it-yourself, DIY, investor. My investment objectives have been to produce moderate portfolio growth (10% annually), low risk (5% maximum drawdown based on monthly returns), no negative yearly returns, and monthly win rates over 70%. These investment objectives are typical of retired investors who cannot afford large drawdowns (no time to recover) but wish to achieve enough portfolio growth to maintain and actually grow their retirement savings while withdrawing around 5% each year for living expenses.
I, myself, fit into this category of retired DIY investors, and I am actively investing in some of the strategies I have presented on SA. I have nothing to promote or gain from these articles; I just want to share my ideas with other DIY investors for their benefit. My reward comes from the feeling of helping others and from the comments that spur me to new and better ideas and strategies.
In this article, a tactical bond strategy is presented that updates monthly and features an asset universe of two high yield classes of bond funds (corporate and municipal) and a GNMA fund. The risk-off asset is cash, i.e. money market. One of these assets is selected each month. This strategy has been spawned off from my earlier Fidelity Quarterly Bond Strategy that I posted on Seeking Alpha at the end of 2015. One of the most significant differences between the earlier strategy and this strategy is the updating frequency; the present strategy updates monthly and can thus react to market conditions more quickly. I personally like monthly updates over quarterly updates.
To show overall feasibility, I have backtested the strategy from 1986 to the present. For backtesting purposes, I have chosen funds that have histories that date back to at least 1986. Thus there are 30 years of backtesting presented. I used the commercially-free Portfolio Visualizer software for the backtests.
The three funds I used in the strategy's universe were:
1. High Yield Municipal Bond Class Asset: MFS Municipal High Yield Bond Fund (MUTF:MMHYX);
2. High Yield Corporate Bond Class Asset: Fidelity Capital and Income Fund (MUTF:FAGIX); and
3. GNMA Class Asset: Vanguard GNMA Fund (MUTF:VFIIX).
The overall correlations between these funds from 04/01/1985 - 10/10/2016 taken from Portfolio Visualizer are shown below. It can be seen that the funds are not well-correlated to each other, as desired in a tactical asset allocation strategy.
Also note that the daily standard deviations are quite low (0.19%, 0.38% and 0.27% respectively) compared to the baseline Vanguard S&P 500 Index Fund (MUTF:VFINX) daily standard deviation (1.14%). This means the strategy will have low volatility.
Of course, correlations can vary over time, and it is more important that correlations are closer to zero or negative during market stress. For instance, many equity assets tend to have high correlation (near 1.0) with each other when there is market stress, so tactical strategies can show poor performance in these times. Another time when low correlation is needed is during times of market churn, when there is no general trend in the market. During these times, tactical strategies can perform poorly, e.g. 2015.
Shown below are correlations between assets during times of market stress (01/01/2001 - 01/01/2003 and 01/01/2008 - 03/01/2009) and a time of market churn (01/01/2015 - 01/01/2016). Tactical strategies need to perform well during these times. It can be seen that the correlations between assets are near zero or negative during these times.
01/01/2001 - 01/01/2003
01/01/2008 - 03/01/2009
01/01/2015 - 01/01/2016
The momentum approach I used was: 1) relative strength ranking of the three funds based on a 3-month look-back period, and 2) a two-month moving average as an absolute filter for the top-ranked fund. So at the end of each month, 3-month total returns (that include dividend distributions) of each fund are determined, and the one with the greatest total return is selected. But the top-ranked fund is only picked if it passes its two-month moving average filter. If the top-ranked fund fails its filter, the money goes to cash for that month.
It turns out that this strategy is quite robust, and the backtest results do not change appreciably if the look-back period is systematically varied between 50 days and 75 days. The moving average filter is equally robust; it can be changed between 10 days and 3-months without any appreciable effect on results.
The backtest results from 1986 - present are shown below. It can be seen that the compounded annual growth rate, CAGR, is 11.9%, the annualized standard deviation is 5.4%, the maximum drawdown, MaxDD, is -5.2%, and the worst year is +0.9% (2005). The monthly win rate is 83.0%. The return-to-risk ratio MAR (CAGR/MaxDD) is 2.3. The correlation to the U.S. market is 0.31.
The maximum drawdown of -5.2% occurs in May 2010 and recovers by October 2010 (in five months). The next highest drawdown of -2.9% occurs from June 2008 to October 2008 and recovers by May 2009 (in seven months). Please note that there is only one drawdown in excess of -2.9%.
An improvement to the strategy is seen when Nuveen High Yield Municipal Bond Fund (MUTF:NHMAX) is substituted for MMHYX and Nuveen High Income Bond Fund (MUTF:FJSIX) is substituted for FAGIX. The improvement in returns using NHMAX, FJSIX and VFIIX is most prominent in more recent years, i.e. 2010 - present. The 2010 - present backtest results with NHMAX, FJSIX and VFIIX are shown below. For comparison, the results for the same backtest with MMHYX, FAGIX and VFIIX are also presented for 2010 - present.
Original Universe (MMHYX, FAGIX, VFIIX); 2010 - present
Improved Universe (NHMAX, FJSIX, VFIIX); 2010 - present
I began investing in this strategy (using the NHMAX, FJSIX and VFIIX version) with my own money at the beginning of 2016. The account has grown 15% YTD, essentially the same as the backtest shows. The volatility has been extremely low, and there has been no drawdown in 2016 (all months have produced positive return).
The only caveat is that the strategy is updated on a monthly basis (although actual trades are generally made much less frequently). So there might be some issues with frequent trading restrictions of the mutual funds. I use Schwab as my broker for this strategy because there are no fees or loads trading Nuveen funds except for a $50 early redemption fee (selling within 90 days). I have reviewed the frequent trading restrictions on the prospectus of each Nuveen fund, and there does not appear to be a frequent trading issue for the Nuveen funds with this strategy.
I feel Schwab might still place their own frequent trading limits on mutual fund trading, so there is still a concern on my part. It is hard to get a true ruling on the matter from online chats or phone calls to Schwab. So far, only one trade has occurred in 2016, so I haven't breached any trade restriction limits at this time. To avoid any issues with trading VFIIX on Schwab, I will instead purchase the ETF version [either Vanguard Mortgage-Backed Securities ETF (NASDAQ:VMBS) or iShares Mortgage-Backed Securities ETF (NYSEARCA:MBB)] when VFIIX is selected in the strategy.
Here is the Portfolio Visualizer link that shows the parameters of the Schwab version of the strategy and the backtest results from 2000 - present.
Disclosure: I am/we are long FJSIX, FAGIX.
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.