Tag Archives: ptmdx

Enhanced Version Of Low Volatility Momentum Strategy

Summary This article continues the work of my previous article on a tactical asset allocation strategy for Schwab or Fidelity platforms using bond mutual funds with very low volatility. The original basket of funds was modified by exchanging one fund for a less volatile fund, and adding a floating-rate loan fund to enhance the strategy when rates are rising. The backtested results show a CAGR of 12.8%, a MaxDD of -2.9%, and a MAR (defining reward/risk) of 4.4. The worst year from 2000 – 2015 had a +5.6% return. Additional details are presented to help understand the practical implementation of the strategy on Schwab or Fidelity platforms. Funds are traded without costs except for a $50 short-term trading fee. The purpose of this article is to present an enhanced version of the Low Volatility Strategy [LVS] that I presented previously (see here ). Based on comments and further study, I have slightly modified the original LVS-1. The -1 designation means one fund is selected each month from a basket of funds. The original LVS-1 had a basket of four mutual funds coming from four different bond classes. Each fund had very low volatility (i.e. daily standard deviations [DSDs] of 0.35% or less) and the funds were mostly non-correlated to each other. A relative strength approach was used in which the funds were ranked based on their total returns over the previous ten trading days. The top-ranked fund was selected at the end of each month unless it failed a 10-day simple moving average [SMA] test, in which case the money went to a safe harbor. The safe harbor was a money market fund. Further details are explained in the previous article. The original basket of funds for application to the Schwab or Fidelity platforms were: Nuveen High Yield Municipal Bond Fund (MUTF: NHMAX ) Principal High Yield Fund (MUTF: CPHYX ) PIMCO Mortgage-Backed Fund (MUTF: PTMDX ) Dreyfus U.S. Treasury Intermediate Term Fund (MUTF: DRGIX ) Changes to Original Basket and Backtest Results After further study, I have replaced DRGIX with the Loomis Sayles Limited Term Government and Agency Fund (MUTF: NEFLX ) because of its reduced risk (reduced DSD that resulted in lower MaxDD). More importantly, I added a floating rate loan fund to the basket in order to improve performance in a rising rate environment. Since I decided to concentrate on the basket of funds for the Schwab and Fidelity platforms, NHMAX limited how far back I could go in a backtest (2000). Thus, I needed a floating rate loan fund with an inception date in 1999 or before. There were three candidates: Oppenheimer Senior Float-Rate Fund (MUTF: OOSAX ): Annualized Return = 4.66%, DSD = 0.18% Invesco Floating-Rate Fund (MUTF: AFRAX ): Annualized Return = 3.62%, DSD = 0.20% Blackrocks Floating Rate Income Portfolio Fund (MUTF: BFRAX ): Annualized Return = 3.76%, DSD = 0.21% OOSAX was selected because it has the highest annualized return and lowest DSD. Thus, the final basket for use on Schwab or Fidelity platforms is: NHMAX, CPHYX, PTMDX, NEFLX and OOSAX. A correlation matrix is shown below, together with annualized returns and various forms of volatility numbers. It can be seen that all funds are noncorrelated except for PTMDX and NEFLX that have a correlation of 0.81. (click to enlarge) Using these funds, LVS-1 was run on Portfolio Visualizer, a commercially-free software package. The backtest was limited to 2000 – 2015 due to the histories of the selected mutual funds. In this article, I am only going to focus on the LVS-1 using mutual funds we will trade on Schwab and Fidelity. However, it should be noted that, in the previous article, this basic strategy was backtested to 1988 using proxies, and good performance and low risk were demonstrated. The results of LVS-1 are shown below, along with results for a buy & hold, equal weight portfolio. Total Return: 2000 – 2015 (click to enlarge) Annual Return (click to enlarge) Tabulated Annual Return (click to enlarge) Drawdown (click to enlarge) Summary Table (click to enlarge) It can be seen that the Compounded Annualized Growth Rate [CAGR] is 12.8%, the standard deviation [SD] is 5.5%, the worst year is +5.4%, and the maximum drawdown [MaxDD] is -2.9%. There are no losing years, and the monthly win rate is 84%. In terms of reward/risk, the MAR (CAGR/MaxDD) is 4.4. This strategy is appropriate for an investor who wants moderate growth and very low risk. Further Thoughts on Implementing LVS-1 on Schwab and Fidelity Platforms The funds that were selected are no load /no fee funds on Schwab and Fidelity. This means the loads are waived, and there are no commission fees. The only fee you will pay is a short-term trading fee of $49.95 if you sell a fund within 90 calendar days on Schwab or within 60 calendar days on Fidelity. So in some instances, you will hold a fund for multiple months, and avoid the short-term trading fee. But most of the time, there will be a charge when you sell a fund. LVS-1 averages about 8 trades per year. That means it will cost about $400 in short-term trading fees per year. For a $100K account, this will come out to 0.4% per year. But there are no other fees. I also looked at the prospectus of each fund pertaining to trading frequency restrictions. All of the funds warn about excessive trading, but they combat excessive trading in different ways. Round-trips are sometimes used to define excessive trading. A round-trip is the buying and selling of one fund in one account. Excessive trading for the mutual funds of interest are: NHMAX: Limited to two round-trips in a 60-day period. CPHYX: Must hold the fund for 30 days before selling. PTMDX: Nothing specific stated. NEFLX: Limited to two round-trips within a rolling 90-day period. OOSAX: 30-day exchange limit. Fund is blocked for 30 days. Thus, there are no limitations that will stop the trading of the LVS-1 strategy as long as we make our trades 30 days apart. This means if we trade on March 1st, our next trade cannot occur until March 31st at the earliest. Conclusion In conclusion, the LVS-1 shows the potential to achieve 12% net growth on average with maximum drawdown (based on monthly returns) of less than 3%. More realistically, this strategy probably has the potential to earn 10% per year with a maximum drawdown of 5%. The monthly win rate should be higher than 80% according to backtesting. As far as I can tell, this strategy should be viable in Schwab or Fidelity accounts as long as the trades are made 30 calendar days apart. To maintain a spacing of 30 days between trades, a schedule is presented in this article . Recently Herbert Haynes has duplicated this strategy and has looked at the effect of trade day on the results. He has shown that trade day is of paramount importance; the only trade days that produce good results are end-of-the-month [EOM] and first day-of-the-month. It is not clear what causes this seasonality of the strategy. Perhaps it is the effect of using funds with large dividends that occur at EOM, or perhaps it is the effect of a short timing period.

Low Volatility Bond Strategy Using Momentum With Short Timing Periods

Summary Most momentum strategies utilize long timing periods, but because of inherently shorter latencies, shorter timing periods would in principle be preferable if whipsaws could be kept within acceptable limits. Shorter timing periods are more likely to work well with low volatility funds. A basket of four funds with daily standard deviations A relative strength tactical strategy named LVS is presented with a 10-day look-back period and 10-day simple moving average cash filter. One fund or two funds are selected each month. Backtested to 1988, the one fund version (LVS-1) has CAGR = 11.6% and MaxDD = -6.6%. The two fund version (LVS-2) has CAGR = 9.4% and MaxDD = -3.4%. LVS implementation is addressed by selecting NTF funds from both Fidelity and Schwab brokerages, and backtesting from 2000 – 2015 with these funds. Monthly win rates over 81% are observed. Tactical momentum strategies rely on look-back periods to establish the ranking of a basket of funds, and then select the best fund(s) to be held each month. The best funds are then filtered by absolute momentum or moving averages. If the funds do not pass their filter, then the money is diverted to a cash fund (usually a money market fund). There are a number of possible momentum strategies that can be used, such as 1) relative strength momentum with cash filters based on absolute momentum. This is commonly called dual momentum; 2) relative strength momentum with cash filters based on moving averages; and 3) a basket of funds with cash filters on each fund based on moving averages. The length of time periods, both for relative strength and for moving averages, is a parameter selected by the developer of a tactical strategy. Developers choose different lengths or combination of lengths based on extensive (or not so extensive) backtesting and/or the research literature. For relative strength, look-back periods ranging from 3-months to 6-months are commonly used. For moving averages, even longer time periods are typically used, e.g. 10-months or 12-months. For moving averages, it is well-known that the optimal moving average can change between asset groups, and can be substantially different under various market conditions. So picking a timing period is not a simple task. In my recent development of momentum strategies (e.g. see here , here , and here ), I have found that shorter timing periods are generally preferred in order to respond quickly to market trends. But short timing periods usually result in whipsaw and poor performance. So the big question is how can we effectively use short timing periods and avoid whipsaw. My answer is that we need to select low volatility funds in our basket of assets, funds with daily standard deviations [DSDs] of 0.35% or less. I arrived at this DSD number after studying what DSD level is needed for effective use of short duration timing periods in tactical strategies. In other words, I determined what DSD level is needed in order for short duration SMAs to produce returns higher than buy & hold, and for maximum drawdowns to be 33% or less of the maximum drawdown of buy & hold. In reality, a DSD of 0.35% is a rather arbitrary number, but it is in the right ball park. To get a feel for DSD numbers for various assets, I have listed DSD numbers for various ETFs and mutual funds from 2009 – 2015: SPDR S&P 500 ETF (NYSEARCA: SPY ): 1.02% Vanguard S&P 500 Index Fund (MUTF: VFINX ): 1.03% PowerShares Nasdaq-100 Index ETF (NASDAQ: QQQ ): 1.11% Vanguard Small Cap Index Fund (MUTF: NAESX ): 1.33% iShares Barclays Long-Term Treasury ETF – 20+ Years (NYSEARCA: TLT ): 0.98% Vanguard Long-Term Treasury Fund – 15+ Years (MUTF: VUSTX ): 0.84% SPDR Barclays Convertible Bond ETF (NYSEARCA: CWB ): 0.69% Vanguard Convertible Securities Fund (MUTF: VCVSX ): 0.59% Vanguard High Yield Corporate Bond Fund (MUTF: VWEHX ): 0.26% Dreyfus U.S. Treasury Intermediate Term Fund (MUTF: DRGIX ): 0.19% Barclays Low Duration Treasury ETF (NYSEARCA: SHY ): 0.06%. All equity assets have DSDs that are substantially greater than 0.35%, and so they can be eliminated from consideration. Some bond assets also have DSD numbers that are too high for use, including long-term treasuries and convertible securities. But there are some bond asset classes that meet the DSD requirement, and also have reasonably high returns. We do not want a fund to have low DSD at the expense of having low annualized return. Thus, a short-term treasury like SHY is not a viable candidate. I have shown previously that a strategy that employs short duration moving averages can be quite effective when used on a low volatility asset. One example is to use the crossover of 3-day and 25-day simple moving averages [SMAs] on the high yield mutual fund VWEHX. If the 3-day SMA is greater than the 25-day SMA, the strategy holds VWEHX. If not, then the strategy is in cash. Total returns and drawdown for this strategy from 2000 – 2014 can be found here . The strategy is quite effective. The DSD of VWEHX over this timespan is 0.25%. In comparison, another high yield asset that is commonly used in tactical bond strategies, SPDR Barclays Capital High Yield Bond ETF (NYSEARCA: JNK ), has a much higher DSD of 0.58% from 2009 – present, and short duration SMAs do not seem to work as well on it as they do for VWEHX. So if my hypothesis is correct, then we need to find a basket of funds that: 1) are non-correlated, 2) have DSDs less than 0.35%, and 3) have relatively high annualized return (> 5%). It turns out that mutual funds are our best option (rather than ETFs) to meet these criteria, and only certain bond classes of mutual funds are suitable. I have found four classes of bond mutual funds that meet the stated criteria. The asset categories are listed below: 1) High yield municipal bond, 2) High yield corporate bond, 3) Mortgage securities, and 4) Intermediate-term treasuries. To show the viability of this approach, I selected four funds that meet the criteria (one from each category). They all have early inception dates. The four funds that I selected are: 1. Oppenheimer Rochester AMT-Free Municipals Fund (MUTF: OPTAX ), 2. Federated High Yield Trust Fund (MUTF: FHYTX ), 3. Fidelity Mortgage Securities Fund (MUTF: FMSFX ), and 4. Dreyfus U.S. Treasury Intermediate Term Fund . The correlations of the funds, together with various forms of standard deviations and annualized returns, are shown below for the timeframe 3/27/1987 to present. These results are taken from Portfolio Visualizer, a commercially-free software package. It can be seen that the funds have relatively low correlation to each other except for FMSFX and DRGIX that have a correlation of 0.74. Notice that all of the funds essentially meet the DSD and annualized return criteria; the only exception is the annualized return of OPTAX that is slightly less than 5%. FHYTX has the highest DSD (0.33%) along with the highest annualized return (7.62%). (click to enlarge) The total returns of the funds from 1999 – present are shown below in a composite figure from StockCharts.com. Please note that the funds complement each other well, i.e. when one or more funds have a downtrend, one or more of the funds have a corresponding uptrend. And, of course, when all funds are trending down, the strategy should put the portfolio money into a money market fund. When all of the funds are trending up, the strategy will select the fund(s) with the greatest momentum. (click to enlarge) I was able to backtest these funds to 1988. I was hoping to go back to 1987, but I could not find an intermediate term treasury fund that had an inception date before 1987. I used the relative strength approach with a SMA as a cash filter. The safe harbor was a money market fund, i.e. CASHX in PV. Because of the low volatility of the funds, a 10-day look-back period for ranking could be used, and a 10-day SMA could be used as a cash filter. These are, obviously, much shorter timing periods than are commonly used in tactical strategies. They can only be used because of the low volatility of the funds. The top-ranked fund was selected at the end of each month. The results of the Low Volatility Strategy selecting one fund each month (LVS-1) are shown below. Portfolio Visualizer [PV] was used to calculate the results. In addition to LVS-1, results are also presented for a buy & hold strategy (updated annually) and for the S&P 500. The S&P 500 was used as a benchmark only because PV did not have a bond benchmark. LVS-1 Using OPTAX, FHYTX, FMSFX, and DRGIX: Total Return, 1988 – 2015 (click to enlarge) LVS-1 Using OPTAX, FHYTX, FMSFX, DRGIX: Annual Returns, 1988 – 2015 (click to enlarge) LVS-1 Using OPTAX, FHYTX, FMSFX, DRGIX: Summary, 1988 – 2015 (click to enlarge) It can be seen that LVS-1 has a Compounded Annual Growth Rate [CAGR] of 11.6%, an annualized Standard Deviation [SD] of 6.3%, a worst year of +0.5%, and a Maximum Drawdown [MaxDD] of -6.6%. In terms of growth/risk, the Sharpe Ratio is 1.2, the Sortino Ratio is 2.7, and MAR (CAGR/MaxDD) is 1.8. Comparison of these numbers to those obtained with a buy & hold strategy and the S&P 500 can be seen in the summary table. Notice that the LVS-1 has the highest CAGR and the lowest MaxDD of the three scenarios. And the MAR of LVS-1 is 1.8 versus 0.5 for buy & hold and 0.2 for the S&P 500. The results for LVS when the two highest-ranked funds are selected (LVS-2) are presented below. From the summary table we see that CAGR drops to 9.4%, but the risk is significantly reduced: SD = 4.5% and MaxDD = -3.4%. The MAR is increased to 2.8. LVS-2 Using OPTAX,FHYTX,FMSFX,DRGIX: Total Return, 1988 – 2015 (click to enlarge) LVS-2 Using OPTAX,FHYTX,FMSFX,DRGIX: Annual Returns, 1988 – 2015 (click to enlarge) LVS-2 Using OPTAX,FHYTX,FMSFX,DRGIX: Summary, 1988 – 2015 (click to enlarge) I will now present a way to practically implement the LVS on Fidelity or Schwab platforms. This requires finding mutual funds with No Transaction Fees [NTF] on Fidelity and Schwab that mimic the funds that have longer historical data (that we have just discussed). I have tried to select NTF funds that do not have any redemption fees, can be traded every 30 days, and have favorable round-trip restrictions per their prospectus. I also tried to find funds that Morningstar rated four stars or higher. The basket is composed of: 1. Nuveen High Yield Municipal Bond Fund (MUTF: NHMAX ), 2. Principal Fields Inc High Yield Fund (MUTF: CPHYX ), 3. PIMCO Mortgage-Back Securities Fund (MUTF: PTMDX ), and 4. Dreyfus U.S. Treasury Intermediate Term Fund . I ran the LVS-1 and LVS-2 with this basket of funds. The backtesting is limited to 2000 – 2015 for this basket because of the limited historical data of NHMAX. The results of LVS-1 are shown below. LVS-1 Using NHMAX, CPHYX, PTMDX, DRGIX: Total Returns, 2000 – 2015 (click to enlarge) LVS-1 Using NHMAX, CPHYX, PTMDX, DRGIX: Annual Returns, 2000 – 2015 (click to enlarge) LVS-1 Using NHMAX, CPHYX, PTMDX, DRGIX: Summary, 2000 – 2015 (click to enlarge) It can be seen that CAGR = 11.7%, SD = 6.0%, Worst Year = +4.0%, MaxDD = -6.5%, and MAR = 1.8. The monthly win rate is 81%; this win rate should be compared with a 60% – 65% monthly win rate for most backtested strategies I have seen. The results compare well with the results using OPTAX, FHYTX, FMSFX, and DRGIX backtested from 1988 – 2015. This strategy is good for an investor who wants high growth and low risk. The results of LVS-2 for NHMAX, CPHYX, PTMDX and DRGIX are presented below. LVS-2 Using NHMAX, CPHYX, PTMDX, DRGIX: Total Returns, 2000 – 2015 (click to enlarge) LVS-2 Using NHMAX, CPHYX, PTMDX, DRGIX: Annual Returns, 2000 – 2015 (click to enlarge) LVS-2 Using NHMAX, CPHYX, PTMDX, DRGIX: Summary, 2000 – 2015 (click to enlarge) It can be seen that CAGR = 9.9%, SD = 4.2%, Worst Year = +3.6%, MaxDD = -2.6%, and MAR = 3.8. The monthly win rate is 83%, an exceptionally high number in a tactical strategy. There are only 33 months with negative returns, out of a total of 190 months. These results also compare well with LVS-2 results using OPTAX, FHYTX, FMSFX, and DRGIX backtested from 1988 – 2015. This strategy is good for an investor who desires moderate growth and very low risk. In summary, a new approach has been conceived that allows the use of short timing periods in tactical strategies without seeing the associated whipsaw effects. To enable the use of short timing periods, each mutual fund in the basket of funds must have a daily standard deviation less than 0.35%. To maximize return, each fund must have adjusted annualized returns over 5%. A tactical strategy named Low Volatility Strategy [LVS] is presented that uses relative strength ranking based on total returns of the last 10 trade days, and a 10-day SMA to filter the top-ranked fund(s). After backtesting LVS to 1988 by using mutual funds with early inception dates, the implementation of the strategy was addressed. NTF mutual funds were selected for use on Fidelity or Schwab platforms. LVS-1 and LVS-2 backtest results from 2000 – 2015 showed monthly win rates over 81%. LVS-1 had a CAGR of 11.7%, a MaxDD of -6.5%, and a MAR of 1.8. LVS-2 had a CAGR of 9.9%, a MaxDD of -2.6%, and a MAR of 3.8. There were no losing years for LVS-1 or LVS-2, with either set of mutual funds.