The Value Line 4 Percent Strategy versus the Value Line 3 Percent Strategy
Another simple market timing strategy is presented in this chapter that signals when to buy to take advantage of market uptrends and then signals when to sell to retain your profits. All the so-called experts on Wall Street will tell you that no strategy can do this consistently. Remember that in the last chapter we reviewed some simple moving-average strategies that do just that.
By using the Value Line 4 percent strategy (hereafter abbreviated VL4%), developed over 30 years ago, or better yet the Value Line 3 percent (VL3%) strategy, you can achieve positive returns. Moreover, both these strategies reduce your overall risk and beat buy-and-hold. The VL4% and VL3% strategies presented in this chapter are for conservative investors with a low risk tolerance. Because the VL4% strategy was thoroughly reviewed in the first edition of this book, along with its long-term track record, this chapter first will focus extensively on the prior and more recent backtests before presenting the better-performing VL3%, so please bear with me as I cover the VL4% first.
BRIEF HISTORY OF VL4% STRATEGY
On June 30, 1961, the Value Line Investment Survey instituted the Value Line Composite Index (VLCI). It represents a weighted index of about 1,700 stocks in 90 industries and thus provides the median performance of all the stocks tracked. The VLCI includes blue-chip, midcap, and small-cap issues. About 80 percent of the stocks are traded on the New York Stock Exchange (NYSE), and 20 percent are traded on the Nasdaq.
This index has a different weighting scheme than the Dow Jones Industrial Average (DJIA) and the Standard & Poor’s (S&P) 500 Index. The DJIA is price-weighted by each of the 30 components (the highest-priced stocks have the most impact on the average), and the S&P 500 Index is market-capitalization-weighted (the number of common shares outstanding multiplied by the stock’s price). In the latter case, the big-cap stocks exert a larger influence on the index. I originally came across the VL4% strategy in Martin Zweig’s Winning on Wall Street (Warner Books, 1986). According to Zweig, Ned Davis of Ned Davis Research developed the “4 percent model” using the VLCI. The model’s goal was to help investors stick with the existing market trend and not get shaken out of the market by small, random day-to-day fluctuations.
HOW THE VL4% WEEKLY STRATEGY WORKS
The VL4% strategy is very simple: If the VLCI rises 4 percent from its last market low (e.g., if it rises from 100.00 to 104.00 or higher), based solely on its weekly Friday closing price, this is a market buy signal, and an investment is made in an appropriate index fund, sector fund, or exchange-traded fund (ETF) of your choice.
If the VLCI declines 4 percent from its last market top (e.g., if it drops from 120.00 to 115.20 or lower), based on a weekly Friday closing price, this is a market sell signal. The proceeds of the sale then can be placed into a money-market account. Alternatively, a more aggressive investor may short the market by using a bear (inverse) ETF. Note that ETFs cannot be shorted in retirement accounts, but you can buy inverse funds in retirement accounts.
Keep in mind that you cannot buy or sell the actual VLCI because it is just an index. The VLCI is tradable as futures or options contracts on the Kansas City Board of Trade (KCBT), but trading in these vehicles is for the more experienced and knowledgeable investor.
The remainder of this chapter focuses on multiple historical tests of the VL4% strategy by different researchers, as well as the latest analysis through October 2010. The earlier backtesting is presented first. If you are interested only in the latest backtests, then skip to page 192 and beyond.
VL4% Weekly Backtest from 1966 to 1988
The VL4% backtested results in this chapter are only theoretical in nature because the trades identified are not executable by an investor owing to the fact that a trading vehicle mirroring the index does not exist. However, backtesting of data against past years is a common and accepted practice used by researchers to determine the viability of specific trading strategies. Of course, backtesting does not guarantee that the same performance that occurred in the past will occur in the future. However, if the same strategy that was tested and worked well in past years still works well today, then that performance says something about the strategy’s longevity and validity. The VL4% fits this paradigm, although it has had its losing periods.
According to Zweig’s book, Ned Davis backtested the VL4% strategy from May 6, 1966, through December 27, 1988. Davis went long on the buy signal and short on the sell signal. His option was to go into cash and wait for the next buy signal, but he decided instead to go short. This decision, as we’ll see, paid off with higher returns. Davis’s results were published in the first (1986) and second (1990) editions of Zweig’s book. The data presented below are from the later edition.
In the 22.7 years of the time period studied by Davis, a onetime initial investment of $10,000 had an ending value for the hypothetical investments made with the VL4% strategy of $233,981, which meant that the investment provided a 14.9 percent annualized return. Buy-and-hold produced a meager return of $17,242, or 2.4 percent annualized. Neither dividends nor transaction costs were factored into this analysis in either case.
The breakdown on the long and short sides of the signals is as follows:
- Longs (buys). Fifty trades were made, averaging 97 days each and producing an annual profit of 16.6 percent. Twenty-six of these trades produced losses averaging –3.9 percent, whereas 24 trades produced gains averaging 14.6 percent.
- Shorts (sells). Fifty-one trades were made, averaging 67 days each and producing an annual profit of 12.5 percent. Twenty-seven of these trades produced losses averaging –3.5 percent, whereas 24 trades produced gains of 9.3 percent.
- Total trades. One hundred and one trades were made, averaging 82 days and producing an annual profit of 14.9 percent. Fifty-three trades showed losses averaging –3.7 percent, whereas 48 trades showed gains averaging 11.9 percent. The ratio of percentage profits to losses was better than 3:1. This is a great example of cutting losses and letting profits run.
In summary, the VL4% strategy, tested by Davis, was a simple timing system that worked well during the 1966 to 1988 time frame. Even though 52 percent of the trades were unprofitable, the average loss of –3.7 percent was far less than the 48 percent of profitable trades, where the average gain was 11.9 percent, a win-toloss percentage ratio of 3.22, which is excellent.
VL4% Weekly Backtest from 1961 to 1992
Nelson Freeburg published an updated analysis of VL4% in his November 30, 1992, newsletter Formula Research. Freeburg backtested the period April 19, 1985, through November 6, 1992, a full four years beyond Davis’s test period.
Freeburg found that the VL4% strategy gained 168 Value Line points over that extended seven-year time frame, whereas the index itself gained only 57 points under a buy-and-hold scenario. On an annualized basis, excluding dividends, the VL4% strategy returned 8.5 percent compared with 3.5 percent for buy-and-hold.
Freeburg even tested the period prior to Davis’s research, going back to July 1961 and testing through mid-1966. During this five-year mostly bullish period, VL4% provided an annual return of 10.7 percent compared with 5.5 percent for buy-and-hold. Thus the VL4% strategy performed well in three different time frames, confirming its consistency and usefulness as a profitable strategy.
Over the entire expanded period from 1961 through 1992, VL4% rose an annualized 13.6 percent compared with about 3 percent for buy-and-hold. This accomplishment is impressive—doubly so because 60 of the 127 buy and sell signals were profitable only 47 percent of the time.
Not satisfied with testing only a fixed 4 percent up or down move on the VLCI, Freeburg tested every percentage point swing from 1 to 8 percent for both the long and short trades to determine if the strategy still would be profitable. After testing 64 parameter sets over 31 years, Freeburg found that all the parameter sets tested exceeded buy-and-hold results by a minimum ratio of 2:1.
The greatest gain in Value Line points occurred with a buy signal set at 4 percent and a sell signal set at 2 percent, that is, a buy if the VLCI rises 4 percent from a bottom and a sell if the VLCI drops 2 percent from a top. This produced 645 Value Line points compared with 149 points for buy-and-hold. Other top combinations were +2 percent and –2 percent, producing 588 points, and the standard strategy of +4 percent and –4 percent, producing 584 points. The worst strategy tested produced 313 points, which still was better than double the buy-and-hold approach. According to Freeburg, this consistent performance over many different percentage point parameters indicates the strength and soundness of the VL4% strategy.
David Penn, a staff writer for Technical Analysis of Stocks & Commodities magazine, wrote an article in the May 2002 issue entitled, “Trends and the 4% Solution.” He reviewed the VL4% weekly strategy from 1985 to 2000. Penn tested the VL4% strategy using the Nasdaq Composite Index and the S&P 500 Index as the investment vehicles. He found, based on the VL4% buy and sell signals, that investing in the Nasdaq Composite Index produced triple the gain of an investment in the S&P 500 Index. Penn did not provide detailed data regarding his analysis in his article. Also, Penn did not provide any performance results or use of the actual VLCI as the index in which to invest. Therefore, his results cannot be compared with Freeburg’s work, which measured only VLCI point changes.
Backtesting Through 2002
I used TradeStation, a popular trading and backtesting platform, to bring the test period current through November 2002 for the first edition of this book. Neither dividends, taxes, nor margin interest were figured into any of the strategies tested with this software.
VL4% Weekly Backtests from 1992 to 2002
I updated the VL4% strategy from December 25, 1992, through November 29, 2002—a period containing both bull and bear markets. The strategy was set up with $100,000 of initial capital. It went long on a VLCI weekly buy signal, and it went short on a weekly sell signal. The funds were never placed in cash because they were always invested long or short the market. During this 10-year time frame, the strategy produced a net profit of $81,998 (see Table 10-1) and a cumulative return of 82 percent. This result compares favorably with a small loss of –1.32 percent for buy-and-hold.
The equity curve in Figure 10-1 shows the investment value after each trade was made. As you can see, there was a big drop after the twenty-first trade, and the fall continued through trade 36. Of these 15 trades, 14 were losers from January 28, 2000, through January 12, 2001. If you looked at a chart of the VLCI during this time frame, then you would see that the VLCI was in a trading range of 980 to 1,180 with numerous up and down moves in price. And each of those trades resulted in a loss. Shortly after you went long on a buy signal, the market reversed down, and a sell-short signal occurred at a lower price, resulting in a loss. Similarly, right after you went short, the market reversed up, and that was offset by a buy signal at a higher price, resulting in a loss. All these whipsaws in price resulted in losses that overcame all the profits built up over the previous years. Actually, the total capital did drop below the original $100,000 investment during this period. From the twenty-first to forty-fifth trades, the VL4% strategy made no progress, as the equity curve shows. This was a disturbing and costly problem. On the positive side, substantial positive annual
returns of 32 percent a year for 2001 and 2002 were impressive because these were down years for the market. And by the end of 2002, the equity curve was at an all-time high.
However, despite this large loss and recovery, the VL4% strategy continued to beat the buy-and-hold benchmark in all the time frames tested by Davis, Freeburg, and me. The cumulative returns for the period 1993 through 2002 may appear to be very low, but remember that we are measuring only the performance of the VLCI itself during this period. If you would have used the VL4% buy and sell signals to trade index funds that tracked the Nasdaq Composite Index, your results would have been better. Recall David Penn’s study earlier, where he tested the Nasdaq Composite Index using the VL4% signals and produced excellent results.
LATEST VL4% WEEKLY BACKTEST 1992–2010
TradeStation was used to evaluate the VL4% weekly strategy from 1992 through 2010, thereby including a portion of the time from the last test ending in 2002 (see Table 10-2). The net profit for the 1992 to 2010 time frame on the original $100,000 investment was $168,958, or an annual rate of return of 18.93 percent. The return on capital was 1,690 percent compared with 382 percent for buy-andhold. The profit factor on each trade was 1.28, just a bit better than breakeven. This test resulted in an 11 percentage point increase in the annual rate of return compared with the prior period ending in 2002. The equity curve (Figure 10-2) had some severe drawdowns during 2000 and during 2010. Table 10-3 provides the annual performance data. Interestingly, the strategy performed very well during the bear market years of 2002 and 2008 but got hit hard in 1995 and especially 2000 (–75 percent drop), the beginning of that threeyear bear market. Thus, overall, this strategy has had mediocre results during the most recent period, taking into account the large drawdowns, but it still outpaced buy-and-hold.
VL4% DAILY BACKTEST FROM 1992 TO 2002
All the published research performed on the VL4% by Davis, Freeburg, and Penn used weekly Value Line data. I have not come across any published findings of research in which daily Value Line data were used. Therefore, in the first edition of the book, I decided to test a VL4% daily strategy using TradeStation.
With the use of daily data, I expected to find many more buy and sell signals with improved performance over the weekly data. This is so because daily data, in theory, should provide faster buy and sell signals and higher annualized returns than weekly data. I tested VL4% daily from December 4, 1992, through November 29, 2002, which resulted in a total profit of $111,234 (as opposed to $81,998 on the weekly test). The annual rate of return for the daily test was 8.14 percent compared with 7.25 percent for the weekly test. The cumulative daily return was 111.23 percent compared with a meager buy-and-hold return of 4.57 percent, so the daily strategy performed better than the weekly strategy by approximately $29,236 during the test period. Unfortunately, the equity curve (not shown) had a big drop in the 2000 time period, where 18 of 20 trades were losers. Thus, even the daily signals could not overcome the numerous whipsaws that produced multiple consecutive losses.
LATEST VL4% DAILY BACKTEST FROM 1992 TO 2010
For the second edition of this book, I ran the same backtest parameters but analyzed the period from 1992 to 2010. This test produced a profit of $460,000, or an annual rate of return of 22.27 percent. This is a return of capital of 4,600 percent compared with 470.5 percent for buy-and-hold. There were 167 trades during this test period, of which 46 percent were profitable. For more details on this test, see Table 10-4.
The annual performance data are shown in Table 10-5. The biggest losses occurred in 1994 (–75.4 percent), 2000 (–48.2 per
cent), and 2009 (–12.7 percent). However, in the bear market years of 2001 (+114.7 percent), 2002 (+63.7 percent), and 2008 (+10.8 percent), there were outstanding gains. Overall, this daily strategy produced 2.7 times the return on capital as the weekly strategy. Therefore, as was the case with the weekly strategy, the volatility and large losses may result in conservative investors not willing to take on the risk of the VL4% strategy altogether. The equity curve (Figure 10-3) illustrates the inconsistent returns over time.
WEEKLY VL3% BACKTEST 1992–2010 OUTPERFORMS VL4% WEEKLY AND DAILY
Based on the mediocre results of the VL4% weekly and daily strategies in more recent years, I decided to determine if a slight variation of the weekly strategy would produce more consistent results, fewer whipsaws, and higher profits. Using TradeStation, I ran an optimization of different values near the 4 percent weekly to see if the performance would be improved for the period 1992–2010. I discovered that simply using a 3 percent value instead of 4 percent resulted in superior results all around. For example, compare the VL4% and VL3% weekly in terms of net profits:
- VL3% $885,293 (long and short) versus VL4% $168,958 (long and short)
- VL3% $647,154 (long only and no shorts) versus VL4% $284,651 (long only and no shorts)
Clearly, using VL3% with long and short trades greatly improved the bottom-line profits compared with VL4%. Table 10-6 provides the details on the VL3% performance results. Table 10-7 provides the annual returns. As you can see, except for the horrific 59.32 percent loss in the brutal 2000 bear market, this strategy had only three other annual losses, all of which were less than 9 percent. And more important, this strategy had tremendous gains in three
bear market years: 2001: +160.89 percent; 2002: +62.53 percent; and 2008: +40.32 percent. Also, in the big bull market year of 2009, this strategy had a 61.36 percent return, beating all the major averages. Moreover, the equity curve (Figure 10-4) is much smoother and steeper than that for the VL4% strategy (refer back to Figure 10-2).
By carefully examining the performance and annual return tables and equity curves for both the VL4% strategies and the VL3% weekly strategy, you can determine for yourself which strategy is superior. Based on my review, it is clear that the VL3% strategy wins hands down in all respects, except that there were 16 more trades for the VL3% weekly strategy compared with the VL4% weekly strategy over the backtest period. The number of
additional trades certainly is manageable based on the low commissions offered by discount brokers and the huge quintuple difference in performance.
Even though the VL4% daily strategy is superior to the VL4% weekly strategy, the much larger number of trades, worse ratio of average wins to average losses, and comparable Sharpe ratio indicate that the daily strategy has some shortcomings. According to Wikipedia, “The Sharpe ratio is a measure of the excess return (or Risk Premium) per unit of risk in an investment asset or a trading strategy. The Sharpe ratio is used to characterize how well the return of an asset compensates the investor for the risk taken; the higher the Sharpe ratio number, the better.”
MTR STOCK-TIMING MODEL (MTR-TM) USES VL4% WEEKLY AS BASIS OF THE MODEL
Additional research on the VL4% weekly model has been conducted by the MTR Investors Group. MTR has worked with and tested the VL4% weekly model and has enhanced it based on additional backtesting. MTR calls its model the MTR timing model (MTR-TM), and it is a mechanical stock market timing model. MTR originally used the VL4% weekly model that issued buy and sell signals based on a Friday-to-Friday percentage change in the VLCI. For example, if the index was up 4 percent from any previous week, MTR went long the market. Likewise, if the index was down 4 percent from any previous week, MTR went short the market. The main issue encountered with this model was that it created too many false (inaccurate) or whipsaw signals. Interestingly, this is the same conclusion that I came to after extensive backtesting.
Accordingly, the MTR-TM was revised with additional parameters that are currently proprietary. This change resulted in better overall net risk-adjusted returns and a higher percentage of accurate trades. The MTR-TM was in testing for over a year until it produced better results for intermediate- and long-term stock market moves than the original unadjusted model. The system was backtested from January 1990 to December 2008. Signals after December 1, 2008, were “out of sample,” and this reflects what would have happened if this model had been live on December 1, 2008. The MTR-TM went live in March of 2009. The model was run from January 1997 to March 2010 and produced 811 percent returns for 72 trades (long and short) compared with a buy-andhold return of 74.49 percent. Table 10-9 provides the data on the backtest.
The vertical lines in Figure 10-5 indicate the buy and sell signal dates. As you can see, they have been accurate. Since it may be a bit difficult to read the exact signal dates, Table 10-10 provides them from January 4, 2008, through September 2, 2010. All of the dates and signals are provided on the MTR Web site (www.MTRIG.com).
At the Web site, the market timing model chart and signals are shown daily, and you can sign up for a nightly e-mail so that you can view them. The signals are provided free of charge; therefore, you should bookmark this site and check it at least weekly for the latest signals if you don’t sign up for the e-mail service. Also, you may want to check out other features of this Web site, including the blog.
TIMING THE MARKET WITH THE VL3% WEEKLY STRATEGY
If you prefer to check on your investments once a week, then use the VL3% weekly strategy. With this approach, you can calculate the percentage change from the previous high or low, as the case may be, once a week (on Friday evening or over the weekend) and then make your investment on Monday if there is a signal change. If at all possible, try to make the investment on Friday afternoon, if you have access to the information, because that timing usually will produce better results than waiting until Monday.
If you are using this strategy for the first time and you are currently in cash, then look for a buy signal when the weekly price goes up by 3 percent from any previous weekly close. Alternatively, a sell signal is generated when the index closes down 3 percent from a prior high. At that point, you can go short or stay in cash if that is your preference.
Once a buy signal is given, you can go long the market with an index fund, an ETF (e.g., a DIA or QQQQ), or an investment vehicle of your choice, as discussed previously. When a sell signal occurs, then you go into cash or buy an inverse (bear) fund—e.g., the ProShares Short QQQQ (PSQ), depending on your risk tolerance. Remember, you cannot short an ETF in a retirement account, but you can buy a bear fund.
To prevent a potentially large loss, you always should use a stop-limit order about 8 to 10 percent below your purchase price depending on your risk tolerance. Of course, it is up to you to select the percentage loss that you are willing to tolerate. Be sure that you cut your losses quickly; otherwise, small losses can turn into big losses.
The use of a mechanical trading strategy, such as VL3% weekly or the MTR-TM, hopefully will take the emotion and uncertainty out of the investing decision. All serious investors should use a systematic investing approach to improve their results. The VL3% weekly strategy requires minimal time to track the signals, provides less risk, and produces a much better return than buy-andhold, but it had some losing years that may be difficult for some investors to stomach. Of course, one approach is to place a stoplimit order after a position is taken based on a VL3% buy signal and then wait for the next signal before taking a new position.
Although the VL4% weekly strategy has been popular over a number of decades, its poor performance during the 2001–2002 bear market and in more recent years has resulted in mediocre results, although it still beats buy-and-hold. Based on the backtested results of the VL3% weekly strategy, investors now have a much better performer to work with. Remember, though, that there is no guarantee that this strategy will continue to outperform going forward. However, the key is to put the odds in your favor. This is what market timing is all about—being on the right side of the market at the right time. Simple strategies do work. Investors who decide to use the VL3% weekly strategy going forward hopefully will perform better than 99 percent of the financial gurus and other investors who have no specific action plan with simple buy and sell rules.