Positive correlations aren't the only way to measure similarities be­tween pairings; negative correlations can be just as useful. In this case, in­stead of a very positive number, we are looking for a highly negative one. Just as on the positive side, the closer the number is to -1, the increas­ingly connected the two currencies movements are, this time in the oppo­site direction. Again we can use the EUR/USD as an example. While we just saw a strong positive correlation with the NZD/USD, the EUR/USD has a very negative relationship with the USD/CHF. Between these two currency pairs, the correlation has been -0.98 over the last year and -0.99 over the past month. This number indicates that these two pairings have a strong propensity to move in opposite directions. Therefore, taking contrary positions on the two pairs could act the same as taking the same po­sition on two highly positive correlated pairs. In this instance going long in one and shorting the other would be almost the same as doubling up on the position, and as a result, would also open your portfolio up to a higher amount of risk. However, deciding to go long or short on both would prob­ably be counterproductive and lead to near zero profit mid losses because the two currency pairs move in opposite directions so if one side of the trade became profitable, the other would usually result in losses.

IMPORTANT FACT ABOUT CORRELATIONS: THEY CHANGE

Anyone who has ever traded the FX market knows that currencies are very dynamic; economic conditions, both sentiment and pricing change every day. Because of this the most important aspect to remember when analyzing currency correlations is that they can easily change over time. The strong correlations that are calculated today might not be the same this time next month. Due to the constant reshaping of the forex environ­ment, it is imperative to keep current if you deride to use this method for trading. For example, over a one-month period that we observed, the correlation between USD/CAD and USD/JPY was 0.06. This is a very low number and would indicate that the pairs do not really share any definitive relationship in their movements. However, if we look at the three-month data for the same time period, the number increases to 0.12 and then to 0.59 for six months and finally to 0.80 for a year. Therefore, for this particular example we can see that there was a blatant recent breakdown in the relationship between these two pairs. What was once a strongly positive association in the long run has almost totally deteriorated over the short term. On the other hand, the USD/CHF and AUD/USD pairing has shown a strengthening trend in the most recent data. The correlation between these two pairs started at -0.78 for the year and edged up to -0.94 for the last month. This suggests that there is an increasing probabil­ity that if one of the trades became profitable the other would incur a loss.

An even more dramatic example of the extent to which those num­bers can change can be found in the GBP/USD and AUD/USD pairing; there was a -0.87 correlation between the two for the yearlong data. How­ever, while these two tended to move in reasonably opposite directions in the long term, between January and March of 2005 they were actually pos­itively correlated with a +0.24 reading. The major events that change the amount and even direction that pairs are correlated are usually associated with major economic happenings such as interest rate changes.

CALCULATING CORRELATIONS YOURSELF

Because correlations have the tendency to shift over lime, the best way to keep current on the direction and strength of your pairings is to calculate them yourself. Although it might seem like a tricky concept, the actual process can be made quite easy. The simplest way to calculate the num­bers is to use Microsoft Excel. In Excel, you can take the currency pairs that you want to derive a correlation from over a specific time period and just use the correlation function. The one-year, six-month, three-month, one-month, and six-month-trailing reading gives the most comprehen­sive view of the similarities and differences between pairs; however, you can decide which or how many of these readings you want to analyze. Breaking down the process step-by-step, we'll find the correlation be­tween the GBP/USD and the USD/CHF. First you'll need to get the pricing data for the two pairings. To keep organized, label one column GBP and the other CHF and then put in the weekly values of these currencies using the last price and pairing them with the USD for whatever time period you want to use. At the bottom of the two columns, go to an empty slot and type in =CORREL. Highlight all of the data in one of the pricing columns, type in a comma, and then do the same thing for the other currency; the number produced is your correlation. Although it is not necessary to up­date your numbers every day, updating them once every couple of weeks or at the very least once a month is generally a good idea.