Over the past few months, I have written a number of articles (here and here) about tactical strategies that employ Vanguard mutual funds traded on a monthly basis. I see the trading of mutual funds (rather than ETFs) as a paradigm shift in tactical asset allocation strategies that trade every month. In recent days I have begun real-money trading, and in the process I have had to deal with implementation issues. This article presents some of the challenges of trading Vanguard mutual funds on a monthly basis, and how I resolved these challenges. Two tactical strategies I have developed using Vanguard mutual funds are also presented: one is a high growth/moderate risk strategy and one is a moderate growth/low risk strategy. These are the strategies I am currently trading with real money.
I have decided to use Vanguard accounts in this testing. I only use mutual funds that have inception dates before 1987, that have no loads/fees, and that only have a 30-day trading restriction. Vanguard spells out what the 30-day restriction encompasses: "After selling a mutual fund, you cannot buy it back within 30 days." Thus, we must have 30 calendar days between trades if we are to successfully trade Vanguard mutual funds.
So my first challenge was to devise a trading schedule that met the 30-day requirement between trades. It turns out that this can be accomplished rather easily. The trading schedule through 2016 is shown in the table below. It can be seen that essentially all trading days occur either on the first trading day of the month (preferred), or the last trading day of the month. There is only one time when a trade has to occur on a day before the last trading day of the month: on December 30, 2015. I have actually assembled a trading schedule through 2019, but I only show the trades days through 2016 below.
Trading Schedule Through 2016
- Friday, 10/30/15: last trading day;
- Monday, 11/30/15: last trading day, 31 days;
- Wednesday, 12/30/15: next to last trading day, 30 days;
- Friday, 1/29/16: last trading day, 30 days;
- Tuesday, 3/1/16: first trading day, 32 days;
- Friday, 4/1/16: first trading day, 31 days;
- Monday, 5/2/16: first trading day, 31 days;
- Wednesday, 6/1/16: first trading day, 30 days;
- Friday, 7/1/16: first trading day, 30 days;
- Monday, 8/1/16: first trading day, 31 days;
- Wednesday, 8/31/16: last trading day, 30 days;
- Friday, 9/30/16: last trading day, 30 days;
- Monday, 10/31/16: last trading day, 31 days;
- Wednesday, 11/30/16: last trading day, 30 days;
- Friday, 12/30/16: last trading day, 30 days.
The second challenge that presents itself is how to determine what funds to select when the trade day arrives. I used the commercially-free Portfolio Visualizer (PV) software to run these calculations, and PV determines the selections based on end-of-the-month data. So there is no issue when a trade needs to be made on the first trade day of the new month. But when fund selections are required on the last trade day of the month, we need a methodology in place to estimate what selections should be made. In the paragraphs that follow, I propose a methodology that has a very high probability to make the correct selections before end-of-the-month data are available. And for those few times when the selections do not agree with end-of-the-month selections by PV, it is probably reasonable to think that it was really a close call anyway, i.e. either selection would have worked or not worked for that month.
Monthly adjusted data (that provide total return) are first obtained for each mutual fund in the basket of funds. So for a 3-month moving average, the adjusted price data at the end of the previous two months is obtained. The latest mutual fund data (on the day before the trade day) is also recorded. One way to obtain an estimate of the end-of-the month price is to just use the price data of the next-to-last trade day as the end-of-month price.
But a better way to estimate the end-of-month price of a mutual fund is to find an ETF that can be used as a proxy for the mutual fund. For instance, the SPDR Barclays Capital Convertible Bond ETF (NYSEARCA:CWB) can be used as a proxy for VCVSX. The ETF can be tracked in real time during the trading day, and its return can be applied to the mutual fund. So if the ETF has increased 0.5% near the close of the market, then the mutual fund is assumed to increase by 0.5% from its previous day price. In this way, the end-of-month mutual fund price can be estimated, and the moving average calculated. Then the mutual fund selection can be determined and appropriate action taken on the trading day (before market close). A spreadsheet, produced by Terry Doherty, has been created that will help facilitate this calculation. It is available to readers upon request in SA private messages.
Two Tactical Strategies Using Vanguard Mutual Funds
My two best Vanguard mutual fund strategies will now be discussed: one (Vanguard High Growth, VHG) for the high-growth, moderate-risk investor and one (Vanguard Capital Preservation, VCP) for the conservative, moderate-growth, low-risk investor.
Vanguard High Growth Strategy
For the VHG Strategy, the objectives are high growth (Compounded Annual Growth Rate [CAGR] greater than 15%) and moderate risk (Standard Deviation [SD] and Maximum Drawdown on a monthly basis [MaxDD] less than 10%). Having no negative annual returns is also an objective. And a final objective is for the MAR Ratio (CAGR/MaxDD) to be greater than 1.5. These objectives are much better than the performance/risk metrics of the Vanguard 60/40 Balanced Index Fund (MUTF:VBINX) that many money managers like to use as a benchmark for a combined equity-bond strategy. The overall metrics of VBINX from its inception in 1992 are: CAGR = 8.06%, SD = 9.05%, MaxDD = -35.06%, MAR = 0.23, and two negative return years.
VHG holds three funds in equal proportions most of the time:
- Vanguard Convertible Securities Fund (MUTF:VCVSX),
- Vanguard High Yield Corporate Fund (MUTF:VWEHX), and
- Vanguard Health Care Fund (MUTF:VGHCX).
Each fund is owned every month except when a fund does not pass its 2-month Exponential Moving Average [EMA]. When the fund fails to pass its filter, the money for that fund goes to the Vanguard Long-Term Treasury Fund (MUTF:VUSTX) for that month.
The backtest results from 1987 through October, 2015 are shown below. The CAGR is 15.04%, the SD is 7.57%, the MaxDD is -9.06%, MAR is 1.55 and there are no negative years. The 2015 YTD return is 12.04%.
VHG: Total Return, 1987 - 2015
VHG: Summary Table, 1987 - 2015
VHG: Annual Returns, 1987 - 2015
Some readers may object to the use of VGHCX as the equity fund in the basket. I selected it because of its superior performance and risk metrics from 1987 to present. Alternatively, the Vanguard Small Cap Equity Fund (MUTF:NAESX) may be substituted for VGHCX. The backtest results (1987 - 2015) using NAESX in place of VGHCX are: CAGR = 13.60%, SD = 8.84%, MaxDD = -10.56%, and MAR = 1.29. Although there are three negative years of return (1987, 1994 and 2002), the returns are essentially zero in those years. So, the strategy with NAESX instead of VGHCX does not breakdown, but the performance and risk are somewhat worse using NAESX instead of VGHCX. For that reason, I decided to use VGHCX as the equity in the VGH Strategy.
The robustness of the VGH Strategy (with VGHCX) is seen in its performance with different moving averages. Comparable performance and risk are seen over a wide range of moving averages, e.g. 2-month Simple Moving Average [SMA], 2-month EMA, 3-month SMA, 3-month EMA, 4-month SMA and 4-month EMA. Here are the overall results (CAGR, MaxDD and MAR) using the various moving averages:
- 2-month SMA: CAGR = 15.44%, MaxDD = -11.94%, MAR = 1.29
- 2-month EMA: CAGR = 15.04%, MaxDD = -9.06%, MAR = 1.66
- 3-month SMA: CAGR = 14.73%, MaxDD = -10.97%, MAR = 1.34
- 3-month EMA: CAGR = 14.10%, MaxDD = -9.37%, MAR = 1.50
- 4-month SMA: CAGR = 13.67%, MaxDD = -11.77%, MAR = 1.16
- 4-month EMA: CAGR = 13.85%, MaxDD = -9.97%, MAR = 1.39
As can be seen, varying the moving averages produces a fairly tight range of CAGRs (13.67% to 15.44%), MaxDDs (-9.06% to -11.94%) and MARs (1.16 to 1.66)
Vanguard Capital Preservation Strategy
The second strategy I will present is called Vanguard Capital Preservation [VCP]. VCP is more conservative than VHG. The objectives of VCP are: 1) CAGR greater than 10%, 2) MaxDD less than 5%, 3) MAR greater than 2.0, and 4) no negative annual returns.
The basket of funds is:
- Vanguard High Yield Tax-Exempt Fund (MUTF:VWAHX);
- Vanguard GNMA Fund (MUTF:VFIIX);
- Fidelity Limited Term Government Fund (MUTF:FFXSX), a substitute for Vanguard Short Term Treasury Fund (MUTF:VFISX). FFXSX was used to enable backtesting to 1987;
- VUSTX; and
The correlation between funds is presented below. It can be seen that the funds are not well-correlated, as desired.
A dual momentum approach is utilized for VCP. The relative momentum at the end of each month is determined using one-month lookback total returns. CASHX (money market) is used as the absolute momentum filter. The three top-ranked funds are selected each month unless they do not pass the absolute momentum filter. If the absolute momentum filter is not passed, then the money for that fund goes to money market. This is in contrast to VUSTX being the safe haven in VHG. For VCP, the money market is the safe haven.
What is somewhat unusual about the VCP Strategy is the short duration lookback period that is used to rank the funds each month. In many dual-momentum strategies, much longer duration lookback periods are used, e.g. 10 months or 12 months. But for some reason, perhaps because less volatile bond mutual funds are used except for VGHCX, a short duration lookback period is optimal. Usually, short lookback periods cause whipsaw in a strategy, but this does not seem to be the case in this strategy.
The backtest results for VCP are shown below:
VCP: Total Return, 1987 - 2015
VCP: Summary Table, 1987 - 2015
VCP: Annual Returns, 1987 - 2015
The CAGR is 11.4%, the SD is 5.61%, and the MaxDD is -4.70%. The Sharpe Ratio is 1.36, the Sortino Ratio is 2.82, and the MAR Ratio is 2.43. The only negative is that the worst annual return is 2015 YTD: +0.6%.
VCP also exhibited robustness in that the lookback period could be changed from 15 trading days to 24 trading days without any significant change in performance or drawdown. VCP actually uses one calendar month (~ 21 trading days on average) for its lookback period.
In summary, this article has presented a way to practically implement tactical strategies using Vanguard mutual funds. To avoid trading frequency penalties, a viable trading schedule has been compiled. In order to decide what selections to make before month-end data are available, a methodology has been presented to estimate month-end data before market close.
The results are presented for two tactical mutual fund strategies that I consider to be the best strategies I have developed in recent months. One strategy (Vanguard High Growth, VHG) should interest investors who want high growth (~15%) with moderate risk (less than 10% SD and MaxDD). For more conservative investors, one strategy (Vanguard Capital Preservation, VCP) produces moderate growth (~11.5%) with very low risk (SD less than 6% and MaxDD less than 5%).
Currently, for November 2015, both strategies are invested in VCVSX, VWEHX, and VGHCX.
Disclosure: I am/we are long VCVSX,VWEHX,VGHCX. 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.