**Introduction**

Over the past three years, I have developed a number of tactical asset allocation strategies that focus on bond mutual funds and bond ETFs. I would like to share this information with you as a gift (no hidden agenda here). My personal goals are to achieve very low volatility and drawdown in my portfolio, and yet realize returns of 8-10% annually (on average). When I talk about low volatility, I am talking about strategies with annualized standard deviations of 3-5% compared to 12-15% for most equity strategies. In regard to low drawdown, I am talking about maximum drawdowns of 3-5% (on a monthly basis) contrasted to 50% for equities. I don't know about you, but for a retired investor like me it would have been quite disheartening to see the value of my portfolio drop by 15% from October - December, 2018. Sure, in this case, equities rebounded and all is now well (maybe), but I'm sure many retired investors were in a state of panic at the end of 2018 if you were investing in equities. I prefer low volatility bond strategies (shown later in this article) that actually returned 1-2% over this October - December, 2018 timeframe.

These bond strategies can be quite productive over all types of market environments; this includes a bond market with increasing long-term treasury rates (although I think such a market is highly unlikely at the current time). Many investors are fearful of bond assets because (they say) we have been in a bull bond market since 1981, but now we are in a bear bond market environment. So (they say) any type of bond strategy at the present time is folly, and backtesting in a bull bond market timeframe has little value (because we are no longer under those conditions). They say we must avoid bond assets at all costs except perhaps very short-term treasuries as a risk-off asset. I'm sure I will hear this type of argument in the comment section below.

But I beg to differ. My tactical bond strategies have done quite well over the past three years when overnight Federal Reserve rates have increased from zero percent to 2.5%. And when long-term rates have increased over the past twenty-five years (it has happened, although infrequently and for rather short time periods), my bond strategies have fared quite well. The reason these bond strategies do well in all types of markets is because the universe of assets includes bond classes that are not well-correlated to each other. That's one key to making bond strategies work in all markets. In contrast to equity classes (where all classes tend to correlate in a bear equity market), bond classes do not all correlate in such circumstances. There is usually at least one class of bond asset that has positive momentum in any given market condition. And, if not, all of my bond strategies feature a safe haven of cash (money market) or short-term treasuries. So for worst case scenario (if all bond classes have negative momentum), my bond strategies will select either a money market fund or short-term treasuries until there is a recovery (i.e. until at least one asset shows positive short-term momentum).

Since my strategies feature low volatility bonds, it is effective to use short duration timing periods in a tactical momentum model without seeing excessive whipsaw. This is another key feature of my bond strategies. When tactical strategies are used for equity assets, there is much higher volatility so longer duration timing periods are required to reduce whipsaw. Typically, timing periods of 10-12 months are used in tactical equity strategies. Such long duration timing periods cause problems in short, quick market downturns like we just experienced from October - December, 2018. Not only did equity strategies with long-duration timing periods remain risk-on through December, but when the recovery started in January, 2019, most went to risk-off.

But with low volatility bond assets, timing periods can be reduced to 1-2 months. With updates either monthly or weekly (depending on the strategy), this means a low volatility bond strategy can rapidly respond to market conditions without being whipsawed. Thus, my low volatility bond strategies had an easy time going to risk-off in November, 2018, and back to risk-on in February, 2019. And in-between, low volatility bond strategies signaled the selection of mortgage-backed/GNMA assets that did quite well in December 2018 and January 2019 (~ +2.5% total return).

Another interesting outcome of low volatility bond strategies concerns the monthly win rate [MWR], i.e. the percentage of months that show positive return. Typically, my bond strategies have monthly win rates of 80-90%; this compares to good equity strategies that typically have MWRs around 60-65%. Because of high MWRs, my bond strategies rarely (if ever) see negative annual returns. Very rarely do low volatility bond strategies experience months with back-to-back losses.

I will now discuss two bond strategies that I have developed by God's grace and invest my own personal retirement money in. More of my bond strategies can be found on my Seeking Alpha [SA] Instablog at this link.

**ELVS**

The first strategy is known as the Enhanced Low Volatility bond Strategy (ELVS). This strategy was first presented in a published SA article in January, 2017, and it can be found here. This SA article should be free to the public (at least I have requested it to be free to SA).

ELVS utilizes bond mutual funds in five uncorrelated classes: high yield corporate bonds, high yield municipal bonds, senior floating-rate bank loans, mortgage-backed/GNMA securities, and short term treasuries. The funds I use in a Schwab account are:

1. Ivy High Income Fund (WHIYX);

2. Nuveen High Yield Municipal Bond Fund (NHMAX);

3. Oppenheimer Senior Floating Rate Bank Loan Fund (OOSAX);

4. Vanguard GNMA Fund (VFIIX); and

5. Goldman Sachs Short Duration Bond Fund (GSSDX).

Here is a link to Portfolio Visualizer that shows the asset correlations for the time period 09/02/1999 - 05/14/2019 based on daily returns. You can see that that these funds are uncorrelated to each other. They are all No Transfer Fee [NTF] funds at Schwab except VFIIX, and VFIIX is still available for purchase. If all NTF funds are desirable, then Voya GNMA Income Fund (LEXNX) can replace VFIIX. The only fees/loads are early redemption fees ($50 if sold within three months) for the NTF funds, and a $50 purchase fee for VFIIX. Typically, there are six transactions per year for ELVS, and this will amount to $300 in fees (on average). These fees are not very significant if you invest $100K or more in ELVS. At Schwab, the funds can be sold and purchased on the same day. I have no experience at Fidelity, but I understand ELVS can be utilized in a Fidelity account if WHIAX is substituted for WHIYX, and the Fidelity GNMA Fund (FGMNX) is substituted for VFIIX.

There are other funds in the same classes that can be substituted for this original ELVS universe. For instance, Diamond Hill Corporate Credit Fund (DSIAX) can be substituted for WHIYX, MainStay MacKay High Yield Municipal Bond Fund (MMHAX) can be substituted for NHMAX, Loomis Sayles Senior Floating Rate and F/I Fund (LSFYX) can be substituted for OOSAX, and Loomis Sayles Ltd Term Government Fund (NEFLX) can be substituted for GSSDX. The fact that we can replace funds with similar funds in the same class is testimony to the overall robustness of the strategy.

ELVS utilizes two timing periods: 15 days and 2 (calendar) months. These timing periods were selected only after extensive robustness testing, i.e. the timing periods were systematically varied to ensure there was little difference in overall backtest results (performance and drawdown). For instance, the 15-day timing period can be varied between 10 days and 20 days without significantly affecting backtest results. I selected the 15-day timing period (halfway between 10 days and 20 days) to ensure overall robustness. The longer duration timing period has similar robustness.

At the end of each month, the five funds are ranked against each other for each timing period based on total return. The overall ranking for each fund is determined by weighting the rankings for each timing period; the weighting is 51% for the 15-day timing period and 49% for the 2-month timing period. Thus, if WHIYX is ranked #1 for the 15-day period and #2 for the 2-month period, then its overall ranking is (0.51 x 1) + (0.49 x 2) = 1.49. The fund with the best overall ranking is the selection for the next month. I make the transaction on the first trading day of the new month, using data at the end-of-the-month [EOM].

What makes the overall ranking calculation somewhat difficult is the EOM distribution that occurs for most bond mutual funds. The EOM distribution is sometimes not included in the adjusted price data until EOM+1 or EOM+2 for many data sources, so until it is included in the data, the returns (based on adjusted data) are not correct. Portfolio Visualizer [PV] can be accurately used for backtesting ELVS, but it cannot be used to determine the next month's selection because some of the EOM distributions may not be included in the adjusted price data until EOM+1 or EOM+2 or even later. So sometimes PV will select one fund at the EOM, and then a different fund on EOM+1 or EOM+2 (once all the EOM distributions are included in the data).

In order to make the correct selection using EOM data, Greg Katai has developed an easy-to-use spreadsheet that can be downloaded for free and is available at this link. There is minimal work required in this spreadsheet to get the next month's selection. The spreadsheet not only shows the selection for the next month, but it also presents the returns and rankings of each fund for each timing period. It is best to not just use the spreadsheet blindly, but rather try to understand the general trend of momentum for all the assets. I sometimes recommend buying two assets instead of one if the differences between the #1 ranked fund and the #2 ranked fund are very small (for diversification purposes).

Another possibility of trading ELVS is to just use the PV signal on the tenth trade day of the month, i.e. at mid-month. PV has recently added a feature that shows the strategy's signal throughout the month. So, at mid-month, all of the EOM distributions are included in the data, and PV's signal is correct. This would eliminate all the headache at the EOM in determining what the correct selection is. I haven't tried this idea yet, but it should be fine and make the selection process much easier, i.e. just look at PV's selection at mid-month.

**Backtest Results for ELVS**

Here are the backtest results for ELVS from PV. The backtest goes from 2000 - present. The Compounded Annual Growth Rate [CAGR] is 12.2%, the Maximum Drawdown [MaxDD] is -2.9%, the number of positive months is 88.4%, and the worst annual return is 4.94% (neglecting 2019). Figures below show Overall Performance, Portfolio Growth, Annual Returns, Monthly Returns, Drawdowns, and 3-Year and 5-Year Rolling Returns. Here is the PV link for ELVS.

Please note that there are 2 1/2 years of out-of-sample testing: 2017, 2018 and 2019 YTD (through April). Returns are 9.25% in 2017, 4.94% in 2018, and 2.89% in 2019 YTD. The 2019 YTD numbers translate to an 8.67% annual return. So even in a general negative bond market environment, ELVS did quite well. MaxDD from 2017 - present is -0.64% based on monthly returns. The annualized volatility is 2.2% from 2017 - present; this compares to an annualized volatility of 11.8% for SPDR S&P 500 ETF (SPY) over this time period.

**MDBS_ETF**

There are many foreign investors who cannot own U.S. mutual funds, so they need a bond strategy that employs bond ETFs. Personally, I prefer bond strategies that employ mutual funds because they have higher returns with lower volatility compared to their ETF counterparts. But a bond strategy with ETFs is also doable.

After trying many different models for bond ETFs, I have settled on a bond strategy called the Modified Defensive Bond Strategy [MDBS_ETF]. Backtesting is limited in scope with bond ETFs because of their short history. So I will first show MDBS_ETF backtests to 2014 utilizing the actual ETFs in the universe of MDBS_ETF, and then I will present longer backtests using mutual fund proxies for the ETFs.

MDBS_ETF uses the relative strength model in PV. The strategy ranks four ETFs:

1. SPDR Blackstone / GSO Senior Loan ETF (SRLN);

2. VanEck Vectors High-Yield Municipal ETF (HYD);

3. Vanguard Mortgage-Backed Securities ETF (VMBS); and

4. Bloomberg Barclays 1-3 Month T-Bill ETF (BIL).

Two timing periods are utilized: 1) a 2-month look-back period (that is assumed to be 42 days), and 2) a 1-month look-back period (that is assumed to be 21 days). The rankings are weighted 51%/49% respectively. This means the longer timing period ranking has the higher weighting. An overall ranking is calculated for each ETF just like I explained in the ELVS description, and the ETF with the highest overall ranking is the winner. Although the original version of MDBS_ETF traded on a monthly schedule, MDBS_ETF now employs a weekly updating schedule, so the winning ETF is bought on the first trading day of the week based on data at end-of-week [EOW]. For most weeks, there is no trade, but the weekly updating schedule allows the strategy to quickly adapt to market conditions (usually without any whipsaw).

There are some brokers that provide commission-free trading of ETFs. Vanguard is one such broker. Schwab has many commission-free ETFs available too. So the cost of trading MDBS_ETF should be negligible. I have also selected ETFs that usually have low friction costs (difference in bid/ask prices); typically there is only a one cent difference between bid and ask prices as long as you trade 30 minutes after market open and 30 minutes before market close. So I usually trade the ETFs during those times and use a market order (for easy trading).

**Backtest Results for MDBS_ETF**

Here are the backtest results for MDBS_ETF from PV. The backtest goes from 2014 - present. The Compounded Annual Growth Rate [CAGR] is 7.4%, the Maximum Drawdown [MaxDD] is -1.1%, the number of positive months is 79.7%, and the worst annual return is 4.91% (neglecting 2019). Figures below show Overall Performance, Portfolio Growth, Annual Returns, Monthly Returns, and Drawdowns. Here is the link to PV for MDBS_ETF.

The out-of-sample testing is more limited for MDBS_ETF. I have been investing real money in MDBS_ETF since about July 2018. The total return from July 2018 - April 2019 is 4.8%, and only one month has shown a loss (October 2018).

To show how MDBS_ETF (and ELVS) compare to equity tactical strategies, I have compared their results from 2015 - present to results of 49 equity strategies tracked by Allocate Smartly. I consider the strategies tracked by Allocate Smartly to be some of the best available to investors. In the figure below, ELVS performance is represented by the blue rectangle, and MDBS_ETF is represented by the blue diamond. The figure comes from Allocate Smartly; I have just superimposed the performance of ELVS and MDBS_ETF on the figure. It can be seen that the two bond strategies provide high return for very low volatility.

###
**Backtest Results for MDBS_ETF Using Mutual Fund Proxies**

Because the backtest of MDBS_ETF is so short, it is always good to substitute similar mutual funds as proxies for the ETFs so the backtest can be extended. Substituting bond mutual funds for bond ETFs is never a perfect match (especially since distributions are handled differently), but it is the best I can do. So, here are the substitutions I made:

1. Eaton Vance Floating-Rate Advantage B Fund (EBFAX) for SRLN;

2. MFS Municipal High Income Fund (MMHYX) for HYD;

3. Vanguard GNMA Fund (VFIIX) for VMBS; and

4. PV's CASHX for BIL.

Here is the PV link to the backtest using the mutual fund proxies. For reference, the backtest results from 2014 - present can be compared with the results using the ETFs. I'm not going to show this comparison (I'll leave it to the reader), but suffice it to say that the overall agreement is very good.

With the mutual fund proxies, we can extend the backtest to 1990. Figures below show Overall Performance, Portfolio Growth, Annual Returns, Monthly Returns, Drawdowns, and 3-Year and 5-Year Rolling Returns. It can be seen that CAGR is 8.9%, the MaxDD is -2.5%, the number of positive months is 81.8%, and the worst annual return is 3.42%.

###
**Conclusion**

I have presented two tactical bond strategies that have been thoroughly backtested and have performed well in live testing. I have various other tactical bond strategies that I invest in; these can be found in detail on my SA Instablog at this link. All strategies have a PV link for ready reference. There is no charge of any kind for this information; I am only trying to help other DIY investors out there.

And here, again, is the link to an article on my SA Instablog that discusses a spreadsheet for three of my strategies. The spreadsheet can be downloaded for free. Each month I post selections on my SA Instablog to seven tactical strategies I follow and invest my own money in. Unfortunately, when I make a post on my Instablog, SA doesn't send an email alert to my followers. So, if you are interested in my monthly selections, you will need to remind yourself to look each month on my SA Instablog.

**Disclosure:** I am/we are long NHMAX,HYD. 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.