In Forex, we talk about going long and short all the time without really considering any difference between them. That’s because in Forex there is not really any “long” or “short”: when you are trading a currency pair, you are really spread trading, always long one currency and short another.
Before we continue, let’s define these terms “long” and “short”.
“Long” simply means you have bought something. If you are long stocks, you have bought stocks.
“Short” means that you have borrowed an asset which you will pay back at the same price it was at when you borrowed it. You hope it will fall, so you can sell it, pay back the original borrowing. And pocket the difference.
In Forex this doesn’t really matter, because you are always buying one currency with another, so you are always long of one currency and short of another one. In trading stocks or commodities it is a little different. You are always either long or short cash against a real asset. Especially in stocks, there is a real difference between long and short. This is because stock markets have a long bias, meaning over any given period of time they have a statistical propensity to rise. Identifying periods of prolonged and continuing falls in the levels of major stock indices is very hard and perhaps even impossible to perform with technical analysis.
The major stock index is the S&P 500, which is an index composed of the 500 major U.S. stocks. It has been in existence since 1957 but it is possible to extrapolate its values for some years before then. Let’s take a look at how this index has performed historically.
S&P 500 Index: Long Bias Demonstration
Let’s imagine first of all that we just bought the Index every week since 1950. This would have produced an average weekly result of 0.18%. This is quite incredible and shows just how resilient the American stock market has been over the past 65 years taken as a whole. It is a clear demonstration of the index’s long bias, suggesting that when you are going short, all other things being equal you have the odds arrayed against you. Of course, we really need to apply a trend-following model to try and get a better idea of the Index’s behavior. Additionally, we will probably get more applicable results if we restrict any back testing to something close to the last 20 years.
So if we look at the S&P 500 Index since 1997, let’s say we buy only at weekly opens where the price is higher than it was both 3 and 6 months ago, and sell when the opposite is the case. This kind of trend-following strategy tends to produce positive results with USD denominated Forex pairs and also with common commodities.
However, when applied to the S&P 500, this strategy can produce positive results on the long side, but only losing results on the short side:
Long weeks: 531 trades, average weekly performance =+0.09%
Short weeks: 227 trades, average weekly performance = -0.06%
No matter which look back periods we use to filter the signals, ANY period of time that we use on the short side produces an average negative result, while just about any period of time we might apply to the long side achieves a positive results.
Which Stock Markets Have a Long Bias?
The first question you might ask yourself at this point is whether all stock indices are like this? We should divide this question into Geography and Sector. For example, if we look at the NASDAQ 100 Index, these are still composed of U.S. stocks, but it is a specific sector of technology stocks. Indices that are composed of specific types of stocks might be more liable to be statistically predictable on the short side.
If we look back through the NASDAQ 100 Index from 1997, we can see that it also has a long bias: in an average week during this period, it rose by 0.12%. Again, we find the same impossibility to construct a profitable momentum strategy on the short side.
Maybe the story is just the amazing resilience of the U.S. stock market. If we tried another key index located in another place geographically, perhaps we can get a good result on the short side.
Geographical Discrimination
What if we looked at a stock market outside the United States? This might give us an example where we can construct a better model for predicting short-term direction profitably on the short side. One good example to use might be the Japanese Nikkei 225 Index, as it is well-known that the Japanese stock market has underperformed over several recent years.
First of all, if we look back from 1997, overall there is actually a SHORT bias here: over the average week, the index fell by 0.28%. So it is no surprise that it is possible to construct a profitable model for the short side here. Let’s see what happens when we apply the 3 month / 6 month test to the Nikkei 225 Index:
Long weeks: 388 trades, average weekly performance =+0.05%
Short weeks: 353 trades, average weekly performance = +0.20%
Here we have a model that is profitable at both ends.
Don’t Use Technical Analysis to Short American Stock Markets
The moral of the story might be that U.S. stock markets have historically been too long to short as part of a trend-following strategy, at least if you are shorting a major index. It might be a better idea to use another judgement as to when to try to get in on a short move. Generally, bear markets fall faster and pass more quickly than bull markets. I think most candlestick analysts would agree with me when I say that a daily chart of the S&P 500 Index is suggestive of a coming fall. However, if you are going to try to take advantage of it, don’t use traditional trend-following techniques, and be prepared to grab profits when the price is falling like a stone!
Note how the Index has struggled to make a new high since March, which is about ten months ago. A few weeks ago a bearish trend line became established across the tops, and over recent weeks we have seen a steeper trend line get established. If this trend line holds, it will soon push the Index down to around the 1900 area. This area has acted as major support twice in recent years, so a strong break below that area might well signify a steeper drop is on its way.
