Trading Strategy: Semiconductors Lead the Stock Market
This trading strategy was inspired by a couple of recent posts from Rob Hanna’s Quantifiable Edges. Rob’s basic idea is that when the semiconductor sector (^SOXX) takes a leadership role in the market, other stocks tend to follow, portending strong returns in the near future.
I’m going to expand on Rob’s idea and share an embarrassingly simple trading strategy with a lot more exposure to the market that is maybe more suitable as a core portfolio strategy.
[logarithmically scaled]
The strategy: go long the S&P 500 at today’s close if the SOXX increases more today (in % terms) than the S&P 500. Close the position at today’s close when it increases less. I told you it was embarrassingly simple. This is just a proof-of-concept, so I’ll assume a frictionless market (no transaction costs, no return on cash, etc.)
The graph above shows the S&P 500 index (blue), the strategy (green), and what happened in the market when the strategy was in cash (red) from mid-1994 to date. Stats below.
That’s a sharp contrast. When the SOXX outperformed the broader market, the stock market performed considerably better the next day.
Three comments about the results above:
First, a look at the graph shows that this strategy underperformed up trending markets, but it also did a VERY good job at protecting against all of the major bear markets of the last 14+ years; and I think that’s the idea’s greatest strength.
Second, this strategy only spent about half of all days in the market. I think there is a lot of value in staying out of the market as much as possible IF you can still accomplish your trading goals. It frees up capital for seeking out other opportunities and reduces the risk of getting caught looking the wrong way in a catastrophic fat-tail day (a’la Oct. of 1987).
And lastly, bear in mind that this test was frictionless – real world considerations could reduce the returns here. But I think that we clearly proved that Rob’s idea has a lot of merit. Would I take a position solely based on SOXX leadership? No. Would I use it as one of many considerations for taking a position? Absolutely.
Happy Trading,
ms
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Stock Market Sectors, Trading Strategies | 9 Comments
Looking back really long ( 50′s or even earlier), each period in the market has had some kind of sector that worked as a leading indicator. SOXX is, apparently, the one that is performing that role today.
However, these leadership role tends to fade away in time. Consumer electronics in the 60′s, as an example, went the same way.
If you look closely at the graph, you will notice that the SOXX leadership curve is reducing is slope. In my opinion, a sign that its leadership is begining to fade.
Unfortunately, most of of these leading indicators are visible only in retrospect. At least that is the case for me. Maybe sharper observers can play the game better.
Eber
Eber – I completely agree. SOXX will eventually fade away (or is fading away now) and will be replaced by something else as a leading sector. Having said that though, I wouldn’t be too quick to dismiss the SOXX continuing predictive-power today. From a return perspective, it has lost some steam recently, but it still appears to be doing a very good job at protecting against market downturns (the most recent one being a good example).
I always enjoy hearing smart commentary like this. Hope to hear more from you in the future Eber.
Happy Trading,
ms
One clarification on the strategy: Do both the SOXX and S&P need to actually increase intraday to trigger a buy in the S&P the next day, or can SOXX increase with S&P decreasing? Can the SOXX and S&P both decrease, with SOXX just decreasing less intraday?
Thank you.
RE to thedude
That’s correct, the SOXX can increase with the S&P decreasing or the SOXX and S&P can both decrease with the SOXX decreasing less. Of course, that’s just how I approached the problem – other combinations might yield better results (don’t think I ran too many iterations of this as it was just a proof of concept).
Hope that helps.
Happy Trading,
ms
Thanks for the clarification. I’m actually new to this kind of analysis, and I was wondering, where does one get the data to do this, and what platform do you use to actually carry out the analysis? Any help would be appreciated.
RE to thedude…sometimes data can be downloaded for free from somewhere like Yahoo (ex. http://finance.yahoo.com/q/hp?s=%5EGSPC). I should note that their data is notoriously a bit buggy, but for most very broad studies like this I feel it’s just fine. Other types of data have to be paid for. I personally like Pinnacle Data (http://www.pinnacledata.com/) because their data is pretty clean, they have a broad selection, and the price is reasonable.
As for analytical platforms, I use a mix of excel, self-created programs mostly using Perl, and sometimes WealthLab (http://www.wealth-lab.com) for strategy backtesting. But don’t let that in any way limit you. There are a lot of other great programs out there that could accomplish the same thing.
In the case of the study above, I used data from Yahoo and crunched the numbers with Excel.
Hope that helps.
Happy Trading (and Researching),
ms
Thanks so much for the help – I’m a total excel geek so I will probably stick to that for the time being. It’s too bad there is no free source of historical options data…