Correlation is one of the more subtle factors which can dramatically impact a trading portfolio. Correlation, or coefficient of correlation, refers to the measure of how two assets’ prices or returns behave against each other over a specified time period. This knowledge enables traders to determine if their positions are truly diversified or merely experiencing the same movement.
Correlation: Correlation is usually measured on a scale from -1 to +1, so a value close to +1 indicates that two assets have historically move in the same direction and a value close to -1 means they have historically moved in the opposite direction of one another. A value close to zero indicates an absence of a linear relationship.
Suppose, for instance, a trader has positions in two currency pairs with different names but, due to the composition of the pairs, ultimately come to be heavily impacted by movement in the US dollar. If there is significant economic news, both might respond in the same way.
Here‘s where correlation comes into play. If you just add more trades you won‘t be reducing risk. The positions are all exposed to the same underlying risk and one market event can damage several trades at once.
Thus, traders ought to consider the relationships among their positions
A frequently seen error made by traders is to think that each individual asset offers diversification. In fact, many trades can be very correlated even if they look completely unsynergies on the trading screen.
Imagine a portfolio of EUR/USD, GBP/USD and AUD/USD. Although they are different currencies, they all have the US dollar and the dollar can move broadly in any direction affecting all three simultaneously. If, for example, the trader takes the same directional view in each position, the total exposure would be much greater than otherwise.
Correlation may also vary with time. Stocks that typically have negligible correlation may show high correlation, for example during times of market panic. Investors tend to respond to the same economic information, to the same interest-rate expectations or changes in risk sentiment; this leads to formerly distinct stocks moving closer together, at least temporarily.
More information available:
Hence this suggests that correlation should not be assumed to be a fixed number. The historical data on correlation is given, but does not imply the relationship will persist.
In portfolio management, traders can use this to watch for large concentrations and ask “if a significant move occurs in a single market factor, how many of my positions can it affect?” True concentrations can be overlooked by a blunt trade count.
Correlation is especially helpful to assess the total risk of a portfolio. Say a trader holds five positions. From the number of trades, the portfolio looks fairly diversified. However if four of those trades tend to be correlated, the unhedged risk could be high.
Here is one example. If you had several currency pairs long, historically having strengthened as the dollar has fallen, then you might have similar underlying exposures. The dollar might spike higher than you expect on some surprise economic news.
That is why position sizing has to account for the correlations between trades. A trader might risk a few percent of their account on a position, but if there are two or more strongly correlated trades in the portfolio the combined worst case break can be substantially higher.
It can exist in the opposite way as well, since negatively and or relatively less correlated positions in the past, may react differently to the same market situation; producing portfolios with a better diversification among its components. This however should not be taken as a quality for one to expect gains and not losses.
Markets are always evolving and correlations can increase, decrease or even change direction.
The objective is not then to find the “portal gun” no perfect combination of negatively correlated assets.
It is also crucial to realize how examination of correlation should take place over different time frames. When looking at correlation via a daily chart, a one day relationship may seem significant, but may not look as clear when assessed over multiple months or within a very short period intraday.
Market conditions may also have a big impact on correlation. When markets are tranquil investors can focus on their unique fundamentals. When a market experiences a heavy shock they will be concerned about matters like interest rates, inflation, liquidity or risk sentiments that together cause several markets to move synchronously.
What does it mean for traders? It should be used as a risk-management tool, not a prediction machine: just because two assets have a strong correlation today does not necessarily implies that they will be correlated tomorrow. Neither does it necessarily means that they cannot be correlated tomorrow.
It may also be convenient to check a portfolio before news developments (central-bank decisions, inflation figures, employment data etc.). If many trades have a common underlying economic factor, the trader may wish to reallocate the overall exposure.
Writing down “trade notes” is another good habit. (I.e. note down why you put on each trade). Over a longer period you can identify exposures that you may not be aware of at the idea stage.
The awesome potential of correlation is that it shifts traders’ paradigm about how we think of risk in a portfolio. This isn‘t just asking if each trade makes sense it‘s analyzing how all of your positions work together.
Having a lot of trades in a portfolio does not mean that it is diversified. If many positions move off of the same currency, commodity, index, economic factor or market feeling, it may act as a big position.
Certainly that isn‘t suggesting traders should run from oncorrelated assets too often. It can work having assets correlated in a planned manner. The key is the understanding how much exposure you take on when adding trades together.
Traders may check their business now and then, assess the relationship of historic period, determine the unity of common risk factors, the total position is suitable for their own risk appetite all at one time. And remind that the correlation statistics are made on history data and might change unexpectedly.
Most important lesson: Counting trades is not equal to measuring diversification. If it is combined with poor security selection, a broader portfolio is sometimes the less diversified one.
Through analyzing correlation of size, market state, and risk limits against.
Comments
Post a Comment