Technical Analysis Review: The Golden Cross

I am by no means an expert in Technical Analysis, but I have studied it with some interest recently.  Much of the money management and academic financial community disregards technical analysis as akin to palm reading or fortune telling.  Since the works of Harry Markowitz, the academic and money management community have accepted the Efficient Market Hypothesis as law.  Anomalies are recognized, but are disregarded as just that.  Anomalies.

Technical analysis incorporates the use of history and charting in attempt to explain why a stock is moving up, down or sideways, and to use that information to glean some insight as to what it may do next.  I won't take sides on the debate, but I'll say that you are doing yourself a disservice by completely disregarding technical analysis.  Isn't it a better idea to study all methods of investing, then determine which method, or combination of methods, makes the most sense to you?

There are numerous terms and rules in technical analysis, most of which I am still not familiar.  There is one that I recently read about called the "Golden Cross" or its counterpart, the "Death Cross."  I did some analysis on the concept.  Here is my definition of the rules, and my analsysis of whether or not it is effective.

Are you familiar with moving averages?  They are a vital part of statistics and used heavily by technical analysts.  A moving average smoothes out the prices movements of a stock or index.  It is calculated by dividing the sum of recent prices by the number of days you are sampling.  You may select the number of days, and in fact, most technical analysts use a variety of moving averages from 15 days to 50 days to 200 days.  The more days you use, the longer term your analysis.  The moving average will trail the price by its very nature.  As prices are moving up, the moving average will be below the price, and when prices are moving down the moving average will land above the current price.  See the chart of AAPL on the first tab of my attached spreadsheet. Download my spreadsheet here.

One of the ways that you use the moving average information is by following a rule called the Golden Cross and the Death Cross.  Here is how it works.  When prices begin a move up, the shorter term moving average will generally begin below the longer term moving average.  So, the 50day MA will be below the 200day MA.  As the price continues its ascent, the 50day MA will cross the 200day MA and continue above it until there is a change the the price action of the stock.  The point at which the shorter term MA moves above the longer term MA is known as the Golden Cross, and is considered a buy point.  Conversely, when the 50day MA crosses the 200day MA to the downside, it is known as the Death Cross, and is considered a sell point.

Take a look at my spreadsheet.  The information is laid out as follows.  The first tab is Apple (AAPL), the pink line is the 200day MA, and the yellow line is the 50day MA.  Currently, the 50day MA is below the 200day MA.  The Death Cross, occured on June 21, 2006.  Interestingly, the stock price is down since that time.

The tab called "AAPL$" is a chart that shows the growth of a hypothetical $10,000 based on whether you used the Golden Cross/Death Cross timing method, or simply bought and held.  You can see that Buy and Hold edged out this timing method.  From January 1, 1998, your $10,000 invested in Apple would have grown to $137,068.97 if you bought and held.  During that same time period, timing the stock by using the Golden Cross/Death Cross method, your position would have grown to $128,925.26.  This also assumes that your money was in money markets while it was out of the stock.  So, you would have had less, but let's remember that you would have avoided some volatile periods, substantially reducing your standard deviation.

The same analysis was done for Krispy Kreme Donuts, Brocade, QQQQs, GE, S&P500, Dow Jones Industrial Average, Russell 2000, and the NASDAQ Index.  Krispy Kreme, Brocade and the Qs did not have a long enough history to begin at 1998.  Instead I began 200 trading days after the earliest price data.

Here are the results:

AAPL (Apple Computers)
Timing:  $128,925.26  Buy and Hold: $137,068.97

KKD (Krispy Kreme Donuts)
Timing:  $11,262.92  Buy and Hold: $5,022.03

BRCD (Brocade)
Timing:  $6840.35  Buy and Hold: $735.08

QQQQ (Nasdaq 100 Trust)
Timing:  $23,315.22  Buy and Hold: $7,480.13

GE (General Electric)
Timing:  $17,336.26  Buy and Hold: $15,901.88

S&P 500 Index
Timing:  $16,629.49  Buy and Hold: $12,881.37

Dow Jones Industrial Average
Timing:  $9,350.17  Buy and Hold: $13,802.72

Russell 2000 (Small Cap Index)
Timing:  $13,319.05  Buy and Hold: $16,042.88

Nasdaq Index
Timing:  $27,082.84  Buy and Hold: $13,112.19

Bottom Line:  Clearly, it does not work all the time and for every period.  Even for some of the ones for which it worked well, there was a period at which it was lagging by a notable amount.  Would you have had the discipline and patience to continue to stick with this strategy?

At first my theory was that it works better for more volatile stocks.  That proved not true.  I think, like any timing mechanism, this works best when applied to stocks that have undergone severe price declines, even if the stock eventually recovers.

During periods of persistent price gains, all timing mechanisms fail.  This is why the generally bullish market of the 80s and 90s convinced everyone that timing the market in any form is for fools.

This one was easy to test.  Feel free to plug in another stock if you would like, and let me know how it works out.  The price data was downloaded from YahooFinance.  If you have difficulty formatting to my spreadsheet, let me know what stock you would like to test via comments.  I'll be more than happy to follow up in the comments.

And most importantly, read my disclaimer:  This is general advice. You should consult with your own financial advisor before making any major financial decisions, including investments or changes to your portfolio.  You, alone, are responsible for any losses or damages that may and will likely result from your financial decisions.

Comments

[Tim MMF]

Great article, it's interesting to see how the "buy and hold" crowd would have done. I need to look into technical analysis more.

[frankyj]

Taking a look at the spread sheet, you don't take into account commissions/taxes (which of course makes it harder to model) but at the same time are a real factor and one of the main reasons technical analysis "fails". Once those real world factors are included, almost any trading rules fail. I'm sure you're aware of that though, just wanted to point that out.

[lamoneyguy]

Frank the FSA, very true. Trading costs, of course, are nearly negligible, depending on the size of the investment. Taxes can take a big chunk, especially if they are short term gains. This is partly why I used longer term MAs.

I think the biggest benefit to technical analysis or any other trading rules is to provide you with another perspective in your decision making process. Often stocks are purchased because an investor believes in the fundamentals of the company. Quickly, the stock turns south, but most investors have difficulty selling at a loss, even when prudent (see charts on BRCD and KKD). Often the fundamentals have not changed, so how do you determine that it is time to sell?

Buy and sell rules give you a realistic process, and adds discipline.

[frankyj]

True. I've heard from more than one person: "I made some investing rules, and if I had just stuck with them, I would have been much better off". Discipline is definitely the key.

Coincidentally, did you notice the dollar-cost averaging story at yahoo? Take a look.

[frankyj]

Maybe if I provided a link :P Dollar Cost Averaging not all it's cracked up to be

[lamoneyguy]

Hmmm... interesting article. I actually think DCA is pretty good for many young invetors. It works because most young people do not have a large amount all at once anyway. Also, it provides a disciplined way to regularly invest money. It does not provide any guidance to ever sell, thus is very popular among mutual fund salespeople.

[Randy]

I was curious what the results might look like on the indexes if the related new 2x Long ETF was used. And what might happen if the related 2x Short ETF was used in place of the T-Bill rate. Since the data isn't available, I assumed the ETF daily change percents are the same for 1x or double for 2x -- which isn't necessarily valid.

For example, where your "Timing" formula was:

=IF(J2144="OUT",K2144*(1+I2143/365),K2144*(1+(F2143-F2144)/F2144))

...I used this for a "2x Long/2x Short" model:

=K2144*(1+IF(J2144="OUT",-2,2)*(F2143-F2144)/F2144)

If my calculations are correct, using the spreadsheet data you provided:

S&P 500 Index
$16,629 Timing
$12,881 B&H
$16,599 Timing/1x Short instead of T-Bill
$20,473 2x Long/2x Short
$22,729 2x Long/T-Bill

QQQQ (Nasdaq 100 Trust)
$27,083 Timing
$13,112 B&H
$30,297 Timing/1x Short instead of T-Bill
$26,098 2x Long/T-Bill
$40,668 2x Long/2x Short

There were four times the 2x Long/2x Short model peaked between $60K and $70K, on wild fluctuations up and down.