In my search for tactical defensive bond strategies in a rising rate environment, I have found that senior floating rate funds such as Oppenheimer Senior Floating Rate Fund (MUTF:OOSAX) provide the best protection. Such funds seem to be better than Treasury inflation-protected funds like Vanguard Inflation-Protected Securities Fund (MUTF:VIPSX). As a reference, Vanguard points out the benefits of senior floating rate funds in a rising rate market in this article.
I personally like simple momentum strategies, the simpler the better. In general, bonds tend to go up when rates are falling and to go down when rates are climbing. In a perfect world, a two-fund tactical bond strategy that switches between a Treasury fund and a non-Treasury bond fund that moves contrary to a Treasury fund should work fine. I have actually found that OOSAX fits this role of a non-Treasury bond fund that has a negative correlation to Treasury funds. More on this in a moment.
As my readers know, I also like low volatility strategies that are safe even in market turmoil. Low volatility strategies lend themselves to short duration momentum look-back periods without producing excessive whipsaw. And short look-back periods enable the Do-It-Yourself [DIY] investor to respond quickly to market trends, i.e. getting into a positive trend early and out of a negative trend early. Of course, it is easy to find short-term, low volatility bond funds, but the trick is to select low volatility funds with enough annual growth to provide moderate portfolio growth in a tactical strategy.
A number of my Vanguard High Growth [VHG] strategies use Vanguard Long Term Treasury Fund (MUTF:VUSTX) as the risk-off asset. Using VUSTX as a risk-off asset has worked well over the past 30 years, but in today's environment, we are concerned about rates increasing and the price of VUSTX decreasing. In such an environment, owning VUSTX as a risk-off asset is probably not prudent. If both equities and long-term treasuries drop at the same, we have no safe haven.
Defensive Bond Strategies: DBS and MDBS
In a previous article shown here, I presented a Defensive Bond Strategy [DBS] that can be used in place of long-term Treasuries as the risk-off asset. The current article presents an improved DBS that should be even better at having a positive return in a rising rate market. I will call this new defensive bond strategy the modified DBS or MDBS.
I selected funds from three classes of bonds for the strategy's universe: 1) senior floating rate, 2) high yield municipal and 3) GNMA. There are various options for NTF funds depending on the broker you use. I will first present three funds for use on a Vanguard account:
1. OOSAX. This is a NTF fund on Vanguard, and the only fee you will pay is a 60-day early redemption fee of $50. If you hold this fund for 60 days or longer and then sell it, there is no fee.
2. Northern High Yield Municipal Fund (MUTF:NHYMX). Likewise, this is also a NTF fund on Vanguard, and the only fee you will pay is a 60-day early redemption fee of $50.
3. Vanguard GNMA Fund (MUTF:VFIIX). This is a Vanguard fund, and there are no fees at all.
Some readers might question why I didn't use all Vanguard funds. Well, first of all, there isn't a senior floating rate fund at Vanguard. Second, Vanguard High Yield Tax Exempt Fund (MUTF:VWAHX) is an option for NHYMX, but the maximum drawdown is significantly reduced with NHYMX compared with VWAHX. The $50 60-day early redemption fee that must be paid if we use NHYMX is a very small price to pay, especially in a $100K account.
Correlations Between Funds in MDBS
The overall correlations based on monthly returns between these three funds using Portfolio Visualizer [PV] are presented below for the 10/01/1999 - 10/31/2016 time frame. I recognize that these overall correlations do not account for the dynamic nature of correlations that actually occur in time, but they at least give us a feel of how the funds interact with each other. VUSTX and Vanguard 500 Index Fund (MUTF:VFINX) are also shown in the table as references.
It can be seen that VFIIX and VUSTX are highly correlated (0.85). It is also noteworthy that the monthly correlation between OOSAX and VFIIX is -0.16, and -0.30 between OOSAX and VUSTX. Hence, in general, OOSAX is negatively correlated with GNMAs and Treasuries. As an example, in 2013 (from May through December), long-term treasury rates increased in anticipation of the end of quantitative easing. Shown below is a comparison of the total return of OOSAX and VUSTX over this time frame. OOSAX increases by 3.7% from 4/29/2013 - 12/31/2013, while VUSX drops by -14.3%. This stock charts figure basically confirms the hypothesis that OOSAX does well when treasury rates rise. Please see the Vanguard reference identified previously for further confirmation.
The monthly correlations of VFIIX and VUSTX to VFINX are negative too from 10/1/1999 - 10/31/2016. Even from 02/29/2008 - 02/28/09 in a bear equity market (not shown), the correlation between VFIIX and VFINX is +0.15, while the correlation between VUSTX and VFINX is +0.20. This indicates that VFIIX generally does well when the equity market is bearish. But there could be times in which Treasuries or GNMA funds might not do well when equity markets are bearish. OOSAX is needed in those times.
Please also note the low levels of Daily Standard Deviation [DSD] for the funds in MDBS: 0.23% or less. This means the MDBS has very low volatility. You should not see sudden drops in your portfolio when you are invested in MDBS.
Momentum Rules for MDBS
The momentum rules of the MDBS are very simple:
1. I used the dual momentum tool in Portfolio Visualizer [PV].
2. For relative momentum, I used a 15-day look-back period. The fund with the highest total return (including monthly distribution at end-of-month [EOM]) over the previous 15 days is the top-ranked fund. I verified that there is very little difference in backtest results if the look-back period is systematically varied between 10 days and 20 days. This speaks to the robustness of the strategy.
3. However, the top-ranked must pass an absolute momentum filter test in order to be purchased. The absolute momentum filter is cash (money market). So if the 15-day total return is greater than the 15-day return of cash (money market), the filter is passed. If not, the money goes to cash.
Backtesting MDBS to 1990 Using Proxies
To enable backtesting to 1990, I needed to substitute proxy funds for OOSAX and NHYMX that had longer histories. I decided to substitute Blackrock Floating Rate Income Fund (MUTF:BFRAX) for OOSAX and MFS® Municipal High Income Fund (MUTF:MMHYX) for NHYMX. PV backtest results from 1990 - present using these proxy funds (BFRAX and NHYMX) together with VFIIX are presented below. MDBS results are labeled Timing Portfolio in the figures. Also shown are the Equal Weight Portfolio [EWP] results and Vanguard Total Bond Market Index Fund (MUTF:VBMFX) results for reference.
It can be seen that MDBS has a Compounded Annual Growth Rate [CAGR] of 7.7%, a Maximum DrawDown [MaxDD] of -2.8%, an annualized Standard Deviation [SD] of 3.2% and a Worst Year of +2.6%. The monthly win rate is 82.3%. There is only one drawdown greater than 2.0% and only three drawdowns greater than -1.25%. The risk-adjusted return ratio MAR [CAGR/MaxDD] is 2.8.
The benefits from the tactical MDBS can be seen when its results are compared to the EWP (rebalanced annually) and buying and holding VBMFX. The MaxDD of MDBS is substantially better than the EWP (-2.8% versus -14.2%), and the CAGR is higher (7.7% versus 5.7%). The CAGR of MDBS is higher than VBMFX (7.7% versus 6.1%) and the MaxDD is about half (-2.8% versus -5.0%). The monthly win rate of MDBS is also substantially higher than VBMFX (82.3% versus 69.0%).
Backtesting DBS to 1990
The PV backtest results from 1990 - present for the original DBS is shown below. The CAGR is substantially less than the MDBS (6.3% versus 7.7%), the Worst Year is substantially less (+0.1% versus +2.6%), and the drawdowns are greater (twelve drawdowns greater than -1.25% versus only three for MDBS). The monthly win rate is also substantially less for DBS (76.4% versus 82.3% for MDBS). Overall, the MDBS shows better performance and fewer drawdowns than DBS. The other advantage of MDBS is that only one fund is selected each month versus three funds for DBS. Having to trade only one fund helps to reduce the number of times we have to pay an early redemption fee.
Backtesting Vanguard MDBS Version to 2000
Next, we will look at PV backtest results for the Vanguard version of MDBS, i.e. using NTF funds of OOSAX, NHYMX and VFIIX. The time frame is only 2000 - present because of the limited history of OOSAX and NHYMX. The results are shown below.
For this Vanguard version, CAGR = 9.6%, SD = 3.8%, MaxDD = -1.9%, and Worst Year = +4.2%. The monthly win rate is 82.2%. There is one drawdown greater than -1.25%. It is interesting that the corresponding numbers using proxies (BFRAX, MMHYX and VFIIX) for the 2000 - present timeframe are: CAGR = 8.1%, SD = 3.5%, MaxDD = 2.8%, and Worst Year = 2.6%. The monthly win rate is 80.2% and there is one drawdown greater than -1.25%. So the use of OOSAX and NHYMX instead of their proxies BFRAX and MMHYX ends up improving performance by 1.5% while maintaining about the same risk. Here is the PV link for the Vanguard version of MDBS.
Backtesting Schwab MDBS Version to 2000
Next, I ran the Schwab version of this strategy. The Schwab NTF funds for this version are OOSAX, Nuveen High Yield Municipal Bond Fund (MUTF:NHMAX) and Loomis Sayles Limited Term Government And Agency Fund (MUTF:NEFLX).
The Schwab version performs much better than the Vanguard version, although the risk is a little higher. The CAGR is 11.4%, the SD is 4.6%, the MaxDD is -2.9% and the Worst Year is +3.4%. The monthly win rate is an amazing 88.1%. This is the highest monthly win rate I have ever seen in a tactical strategy. There are four drawdowns of -1.25% or more. In my opinion, these results are so good that this version can be used as a stand-alone strategy. Here is the PV link for the Schwab version of MDBS. The better performance of the Schwab version is mainly caused by the superior performance of NHMAX.
Effectiveness of DBS and MDBS as Risk-off Strategies
1. BTS with VUSTX as Risk-off Asset
To show how MDBS can be used as a risk-off asset in place of VUSTX, I ran the Balanced Tactical Strategy [BTS] that is discussed in the original DBS article. The BTS was developed to use VUSTX as the risk-off asset, and it has performed well over the past 30 years. However, I am concerned that BTS's performance in a rising rate environment might be disastrous when the equity market and Treasury markets are bearish at the same time. Here are the PV backtest results from the 1990 to present timeframe using VUSTX as the risk-off asset.
BTS shows promising results with VUSTX as the risk-off asset. There is moderate growth (CAGR = 11.2%) and limited risk (MaxDD = -7.1%). Only one year shows negative growth (1994 = -1.6%); all other years have positive return. The monthly win rate is 70.2% and there are 18 drawdowns in excess of 2.0%. The MAR is 1.6.
However, having VUSTX as the risk-off asset seems problematic in today's market.
2. BTS with DBS as Risk-off Strategy
If we replace the risk-off asset with the original DBS, there is improvement in risk and a corresponding drop-off in growth. Here are the overall PV results for BTS from 1990 - present with DBS as the risk-off strategy.
Using DBS as the risk-off strategy instead of VUSTX definitely has its benefits. The SD is 4.3% versus 6.2% when VUSTX is the risk-off asset. The MaxDD is improved from -7.1% with VUSTX to -3.1% with DBS. The drawdowns greater than 2.0% decreases from 18 with VUSTX to 10 with DBS. The monthly win rate goes from 70.2% with VUSTX to 72.0% with DBS.
But the improved risk with DBS comes with a price: lower CAGR. The CAGR with VUSTX is 11.2%; the CAGR with DBS is 8.7%. So there is a penalty of 2.5% reduction in CAGR.
3. BTS with MDBS as Risk-off Strategy
To complete our study of defensive bond strategies, I ran the BTS with MDBS (using proxy funds). The PV backtest results from 1990 - present are presented below.
These results show that there are some benefits using MDBS instead of DBS as the risk-off strategy for BTS. The CAGR is increased somewhat (9.0% versus 8.7%) and the MaxDD is improved a little (-3.1% with DBS to -2.9% with MDBS). The monthly win rate significantly increases from 72.0% with DBS to 77.0% with MDBS. The number of drawdowns greater than 2.0% improves from 10 with DBS to 9 with MDBS. So, MDBS seems to produce better performance with lower risk than the original DBS. But the overall improvement is relatively minor.
This article presents a modified defensive bond strategy [MDBS] to be used as a risk-off strategy in place of long-term treasuries like VUSTX. It has been shown that MDBS, by itself, is slightly better in performance and risk than the original DBS, plus it has good potential to counteract the negative effects of rising rates going forward that DBS lacks. In a Balanced Tactical Strategy [BTS], MDBS as the risk-off component provides good risk control compared to VUSTX as the risk-off asset. However, there is a price to pay for improved risk control: loss of annual growth.
The results of both Vanguard and Schwab versions of MDBS are presented. The Schwab version has such good performance that it can be considered as a stand-alone tactical strategy.
Disclosure: I am/we are long OOSAX. 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.