Popular Strategy to Yield Buy Signal
It has become a popular strategy over the past decade, and it does have some merit, but there are some serious pitfalls.
The strategy is to buy the SP500 when prices close above the simple 10 month moving average; the sell signal is a monthly closing price below the simple 10 month moving average. In all likelihood as the month nears its finale, the SP500 will close above its simple 10 month moving average thus triggering a buy signal.
This strategy has been popularized by Mebane Faber of the World Beta blog and others around the internet have taken up the mantle, as well. I would characterize the strategy as elegantly simple. But simple can be good too. The strategy does work and it works across multiple asset classes. It is easy to implement as you only have to look at monthly charts and ask yourself one question: where is price relative to the simple 10 month moving average? The strategy will keep you in bull markets, and by moving to cash when prices are below the simple 10 month moving average, you avoid the punishing draw downs (and higher volatility) of a bear market.
Ah, if only it were so easy. So what are the pitfalls?
First some numbers. Since 1960, this strategy generated 1565 SP500 points. Buy and hold SP500 netted 1265 points. The strategy generated 36 round trip trades over the years with 67% of the trades being winners. The average winning trade was almost 4 times as large as the average losing trade. The average trade lasted 13 months; the average winning trade lasted 17 months with the average losing trade going on for only 4 months. No trade loss more than 11% but only two trades had individual trade draw downs exceeding 15% (1990 and 1987). This means that 34 out of 36 trades had individual trade draw downs less than 8%. This is pretty impressive stuff. So where is the rub?
One concern I have with this strategy is how it made its money. There was one trade — from 1994 to 1998 — that was responsible for almost 33% of the points earned from this strategy. While we can argue that’s the luck of trading or how bull markets go, the reality is that one trade in 52 years was a statistical outlier. My question: do we want to have a strategy that is dependent upon capturing a statistical outlier to be responsible for that strategy’s success? The answer is “no”. We would prefer to have strategies that earn their gains on a consistent basis. If this one mega trade did not occur, it is very likely that this strategy would have underperformed buy and hold.
Another concern for this strategy is that it is not particularly efficient. By this I mean, individual trades can give up significant gains before being closed out. In other words, sell signals typically occur long after the market has peaked. This can be seen in the Maximum Favorable Excursion (MFE) graph. See figure 1. MFE, measures in percentage terms, how far a trade moves in your favor before being closed out for a loss or a win. So take the individual trade inside the blue box in figure 1. This trade ran up 39.45% (x axis) before being closed out for 31.05% (y axis) winner. We know this was a winning trade because the caret is green. The trade has an MFE of 32.69% but loss 8.40% before a sell signal was generated. Looking at the trades inside the oval, we note that they all ran up near 30% but gave up one-third of those gains before being closed out. Even the mega 1994-1998 trade (see orange arrow on chart) gave up over 50% in gains before being closed out. Ouch!
Figure 1. MFE Graph
In summary, there is a lot to like about this simple, easy to implement strategy. However, under the surface there are several “issues” that might make it difficult for investors to utilize and follow. Such is trading. The important lesson from this exercise is to avoid the big losses, and that is what this strategy does well.
What does this buy signal mean for the markets and economy? It probably is a signal that a recession has been averted (for now). It would be unlikely to have the economy in recession with prices on the SP500 above its simple 10 month moving average.
Category: Technical Analysis
Comments (10)
Trackback URL | Comments RSS Feed
Sites That Link to this Post
- Sunday links: meddling in the results | Abnormal Returns | January 29, 2012
- Top clicks this week on Abnormal Returns | Abnormal Returns | February 5, 2012
You are wrong about this.. this is not an outlier.. the strategy is designed to capture such positive outliers.. whenevr such an event happens again, it would be captured by the strategy.. this is the beauty of such trend following systems.. not a drawback
But if a move of this magnitude doesn’t occur again within your trading lifetime, your overall performance won’t likely exceed buy and hold. That’s the author’s point. If that’s OK with you, fine.
What I would quibble about is the claim that the S&P moving above the 10 month avg means that a recession has been averted. If federal monetary and fiscal policy were the same today as it was up until about 2000, this might be a good call. But with the government actively involved in the markets (by pumping printing press money into the banking system) this whole rally could be built on nothing more than “If businesses are too scared to borrow this cash for expansion, what the hell do we do with it?”
Johninohio
I agree with everything you say, but in Q4, 2010 the economic data looked like a recession, but the Fed announced QE2 adn the market took off; the economic data was consistent with recession for 5 more months after the SP500 closed above its simple 10 month moving; price is an excellent abeit short term leading indicator
Piyush
I am not wrong! Once every 52 years is an outlier! You cannot wait for that single event to occur once every 52 years
plyush modi: Please explain. If the sell signal in the outlier exception was based on the price declining below the 200 DMA would not the “give up of points” from the peak profit potential not matter – in that if the strategy to but and sell at breaks above and below the 200DMA was not changed – (which would negate the strategy) how would this simple strategy avoid such? Am I missing something?
Re the article:I would think that USING the 200 DMA strategy could be used assuming a modification: This would use the use of 200 DMA’s on market and sector/industry indexes to create ONE signal and the 200 DMA of an individual stock within this market/sector/industry as a “confirming” second signal.
Or one could use the 50 DMA as a confirming signal. any combination of using such signals could enable an investor take initial action on the one signal, add or liquidate a second position on the next signal and complete the position on the third confirming signal.
Example
ABC stock in industry XYZ
each has 50 and 200 day moving averages
based on fundaments, one wants to buy ABC.
INITIALLY, the investor buys a position based on the number of “buy indicators” defined as ABC and/or XYZ being above their 50 or 200 day MAL. and adds (25%) positions if and when the final (of four) bullish (above the remaining) 50 or 200 day MAL.
Just a thought/concept to discuss. I use this but my problem is that I do not always act on it. But that does not mean that such combination of 4 indicators might not improve the statistical odds of improving the profit outcome with less opportunity loss when the stock moves up a lot above it’s 50 and 200 day MAL.
One buys an initial (25%) position if and when ABC and or XYZ moves above their 50 OR 200 day MA.
The investor adds to the ABC position if and when ABC or XYZ moves above the 50 or 200 DMA.
Piyush:
Remember we are following the signals strictly; wait until the end of the month to take buy and sell signals….you haven’t even traded the strategy and you are already making changes
I don’t care what you do; the purpose of backtesting is to create one’s expectations
I didnt talk about making any changes. you are probably referring to romeo foxtrot here.
In general I have to point that this buying above 10 month moving averages and 200 dma averages typically doesn’t beat buy and hold by much. And this is said even by people who talk and recommend such strategies. Where it does work very well is in reducing drawdowns and volatility in returns.
@ foxtrot – such strategies are only traded on the long side and not on short side