Seeking Alpha
Profile| Send Message| ()  

In the first article, we predefined a fix number of currencies (Australian, Canadian, and American dollar, Swiss franc, and Bristish pound), and we calculated a reasonable portfolio to hedge against euro. In this second article we want to determine why these currencies and why we started using 20% of each one when some of them behave approximately the same. For instance, the Australian and Canadian dollar have a high correlation and if we use them separately we are overweighting their importance in our basket.

Sometimes it´s better to show not-so-easy concepts through simple examples. Let´s imagine that we live in U.K. Our main currency is the pound, then. However we travel a lot and are quite flexible to live in other countries. Besides we don´t trust the pound that much, so as to have all our wealth in this currency. We keep on imaging that we have a salary of 2000 pounds a month, but have 500.000 pounds equivalent to invest. Again, as in the first article, we may use certain ETFs, such as FXF, FXA, FXC, to gain exposure to other currencies.

(Click to enlarge)

First thing we need to know is if certain currencies are highly correlated. If this is the case, they behave as one. To check out correlation patterns there are many calculators like this one from Mataf Forex. As we are long term investors, we select the currencies we want to display in the table, 200 in "Number of periods" and we just look at the weekly table, the last one. In general, on one side we have USD and the rest on the other side. So, if we choose EURUSD, AUDUSD, USDCAD, USDCHF, and GBPUSD, we will see that AUDUSD and USDCAD are highly correlated (-91.8, negative because in the first one dollar is in the denominator and in the second one in the numerator), so it makes sense to consider Australian and Canadian dollar as one currency. Amazingly, both are also highly correlated with CHF.

Second, we check if these correlations are stable. If we pivot our currencies around the pound, Aussie-Loonie-Swissy have a correlation greater than 80%. We check this correlation out with different periods and somehow it´s valid. So let us consider these 3 currencies as one (1/3), euro (1/3), and U.S. dollar (1/3).So instead of using 20% as we have shown in the first article, we use 11% for Aussie, 11% for Loonie and 11% for Swissy, 33% for U.S. dollar, and 34% for euro

As a result, with the data from the article I, our portfolio would be: CHF>7%,AUD>13%, CAD>11%, EUR>33%, and USD>36% taking into account that we have supposed that our home currency is pound.

Source: How To Build A Currency Basket II